Likely size of the French BSE epidemic - Epidemiological analysis helps in evaluating the potential risks of eating French beef.
Likely size of the French BSE epidemic - Epidemiological analysis helps in evaluating the potential risks of eating French beef.
复制标题
DOI：
10.1038/35048666
复制
发表时间：
2000-12-14
期刊：
影响因子：
64.8
通讯作者：
Donnelly, CA
中科院分区：
文献类型：
--
作者：
Donnelly, CA
到目前为止，英国报告的牛海绵状脑病病例最多（177,490头），但其他欧洲国家也报告了本土出生的病例（爱尔兰共和国524例；葡萄牙447例；瑞士350例；法国143例；比利时、丹麦、德国、列支敦士登、卢森堡、荷兰和西班牙，均少于20例）。在这里，我分析了来自法国的疯牛病发病率数据，以估计那里的流行病的规模和进程，并发现自1987年中期以来，至少有1200头法国牛感染了导致疯牛病的病原。这表明，即使在对法国疯牛病感染发生率最乐观的假设下，2000年法国估计将有49只受感染的动物被屠宰供人食用。图1法国疯牛病发病率，按临床发病年份划分的疯牛病病例发病率（截至2000年11月17日确诊和报告；数据仅包括法国出生的牛）。b，按队列估计的疯牛病感染发生率（95%置信区间），允许漏报（红色）和假设完整报告（蓝色）。每年出生队列的定义如下：例如，1992年队列包括1991年7月1日至1992年6月30日之间出生的牛。疯牛病的年龄特异性发病率通过出生队列建模，表示di （a,c）为年龄和出生队列，asI(a,c) = Δ（a+c）S(a)N(c)f(a)，其中漏报建模为报告疯牛病病例的时间依赖概率，Δ(a+c) = exp(α+β(99−a−c))/[1 + exp(α+β(99−a−c))] fora+c= 99和Δ(a+c) = 1 fora+c= 100（参考文献）。S(a)是动物存活到18岁的概率，N(c)是感染疯牛病病原的动物数量，f(a)是感染动物在18岁时出现临床发病的概率（假设所有动物都存活到18岁）。假设出生队列数据的疯牛病年龄特异性发病率来自泊松分布，使用最大似然方法估计参数α和β。发病年龄分布f(a)（感染年龄和潜伏期分布的卷积）和生存分布S(a)均取自英国疯牛病流行的反计算分析得到的有效参数估计值和英国牛的生存分布。在一组平行分析中，每个队列（I(a,c) fora+c= 100）在最后一个年龄类别中观察到的发病率增加了21%（根据1999年11月至2000年11月期间inI(a,c) fora+c= 99的变化估计），以调整确认延迟，以及尚未发生的临床病例的发病，例如，1994年队列中的动物可能在2001年初6岁时出现临床发病。由于这一修正系数，如果考虑漏报的情况，2000年的感染总数和被屠宰的受感染动物数量大约增加19%（如果报告完整，则增加13%）。
The United Kingdom has reported the largest number of cases so far of bovine spongiform encephalopathy (177,490 animals), but native-born cases have been reported in other European countries as well (Republic of Ireland, 524; Portugal, 447; Switzerland, 350; France, 143; Belgium, Denmark, Germany, Liechtenstein, Luxembourg, the Netherlands and Spain, each fewer than 20 cases). Here I analyse BSE-incidence data from France to estimate the size and course of the epidemic there and find that at least 1,200 French cattle have been infected with the aetiological agent that causes BSE since mid-1987. This suggests that, even under the most optimistic assumptions regarding the incidence of BSE infection in France, an estimated 49 infected animals will have been slaughtered for human consumption in France during 2000.Figure 1BSE incidence in Francea,BSE case incidence by year of clinical onset (confirmed and reported by 17 November 2000; data include only French-born cattle).b,Estimated BSE infection incidence by cohort (with 95% confidence intervals) allowing for under-reporting (red) and assuming complete reporting (blue). Annual birth cohorts were defined so that, for example, the 1992 cohort consists of cattle born between 1 July 1991 and 30 June 1992. The age-specific incidence of BSE was modelled by birth cohort, denotedI(a,c) for ageayears and birth cohortc, asI(a,c) = Δ(a+c)S(a)N(c)f(a), where under-reporting is modelled as a time-dependent probability that a BSE case is reported, Δ(a+c) = exp(α+β(99 −a−c))/[1 + exp(α + β(99 −a−c))] fora+c= 99 and Δ(a+c) = 1 fora+c= 100 (ref. ).S(a) is the probability that an animal survives to agea,N(c) is the number of animals in cohortcinfected with the aetiological agent of BSE, andf(a) is the probability (assuming all animals survive to age 18 yr) that an infected animal will experience clinical onset at agea. The parameters α and β were estimated using maximum likelihood assuming that the age-specific incidence of BSE by birth-cohort data arose from a Poisson distribution. The age-at-onset distribution,f(a), (the convolution of the age-at-infection and incubation-period distributions) and the survival distribution,S(a), were taken from validated parameter estimates obtained from back- calculation analysis of the British BSE epidemic,,,and survival distribution of British cattle,. In a parallel set of analyses, the observed incidence in the last age category for each cohort (I(a,c) fora+c= 100) was boosted by 21% (estimated from the change inI(a,c) fora+c= 99 between November 1999 and November 2000) to adjust for confirmation delay, as well as for the onset of clinical cases that have yet to occur because, for example, animals in the 1994 cohort could experience clinical onset at age 6 yr in early 2001. As a result of this correction factor, total infections and the number of infected animals slaughtered in 2000 were roughly 19% greater if allowance was made for under-reporting (and 13% greater assuming complete reporting).