Boelaert F, Walravens K, Biront P, Vermeersch J P, Berkvens D, Godfroid J
Veterinary and Agrochemical Research Center (V.A.R.), Groeselenberg 99, 1180, Ukkel, Belgium.
Vet Microbiol. 2000 Dec 20;77(3-4):269-81. doi: 10.1016/s0378-1135(00)00312-6.
The national bovine paratuberculosis (PTB) seroprevalence (apparent prevalence) in the Belgian cattle population was determined by a serological survey that was conducted from December 1997 to March 1998. In a random sample of herds (N=556, 9.5%), all adult cattle of 24 months of age or older (N=13,317, 0.4%) were tested for the presence of antibodies using a commercially available absorbed ELISA test kit. The PTB median within-herd seroprevalence (proportion of detected animals within the seropositive herds) and the PTB individual-animal seroprevalence (proportion of detected animals) were, respectively, 2.9% (quartiles=1.6-5.6) and 0.87% (95% confidence interval (CI)=0.71-1.03). The PTB herd seroprevalence (proportion of detected herds) was 18% (95% CI=14-21). Assuming a test sensitivity and specificity of 45 and 99% [Sweeney et al., 1995. J. Vet. Diagn. Invest. 7 (4), 488; Sockett et al., 1992. J. Clin. Microbiol. 30 (5), 1134], respectively, the median true within-herd prevalence and the true individual-animal were estimated to be 7 and 2%, respectively. The true herd prevalence of Mycobacterium paratuberculosis infection was first estimated according to currently accepted methodology. This calculation revealed that the specificity of the used test has a dramatic effect on the estimation; assuming a test sensitivity of 45% and a true within-herd prevalence of 7%, the true herd prevalence estimation decreased from 36 to 0.8% if the test specificity decreased from 99. 9 to 99%, respectively. This sensitivity analysis showed that the practical limits of the accuracy of the used screening test jeopardize the estimation of the true herd prevalence within reasonable confidence limits, because the within-herd PTB true prevalence was low. For this reason we augmented the herd specificity for herds with larger adult herd size (>5). This was done by increasing the cut-off number of positive cattle required (>/=2) to classify a herd truly positive and including herds with one positive test result if there was historical evidence of PTB (previous diagnosis and/or clinical signs). This approach resulted in an estimated true herd prevalence of M. paratuberculosis infection of 6%. The true herd prevalence for dairy, mixed and beef herds was, respectively, 10, 11 and 3%.
1997年12月至1998年3月进行的一项血清学调查确定了比利时牛群中牛副结核病（PTB）的全国血清流行率（表观流行率）。在一个随机抽取的牛群样本（N = 556，占9.5%）中，对所有24月龄及以上的成年牛（N = 13317，占0.4%）使用市售的吸收酶联免疫吸附测定试剂盒检测抗体的存在情况。牛群内PTB血清流行率中位数（血清阳性牛群中检测到的动物比例）和个体动物PTB血清流行率（检测到的动物比例）分别为2.9%（四分位数=1.6 - 5.6）和0.87%（95%置信区间（CI）=0.71 - 1.03）。PTB牛群血清流行率（检测到的牛群比例）为18%（95% CI = 14 - 21）。假设检测敏感性和特异性分别为45%和99%[斯威尼等人，1995年。《兽医诊断调查杂志》7（4），488；索基特等人，1992年。《临床微生物学杂志》30（5），1134]，估计牛群内真实流行率中位数和真实个体动物流行率分别为7%和2%。首先根据目前公认的方法估计副结核分枝杆菌感染的真实牛群流行率。该计算表明，所用检测方法的特异性对估计有显著影响；假设检测敏感性为45%，牛群内真实流行率为7%，如果检测特异性分别从99.9%降至99%，则真实牛群流行率估计值从36%降至0.8%。该敏感性分析表明，由于牛群内PTB真实流行率较低，所用筛查检测方法准确性的实际限制危及在合理置信限内对真实牛群流行率的估计。因此，我们提高了成年牛群规模较大（>5）的牛群的特异性。通过增加将一个牛群判定为真正阳性所需的阳性牛的临界数量（≥2）来实现这一点，如果有PTB的历史证据（先前诊断和/或临床症状），则将有一个阳性检测结果的牛群也包括在内。这种方法导致副结核分枝杆菌感染的估计真实牛群流行率为6%。奶牛群、混合牛群和肉牛群的真实牛群流行率分别为10%、11%和3%。
