A relevant long-term impact of the circulation of a potentially contaminated vaccine on the distribution of scrapie in Italy. Results from a retrospective cohort study

Table 2 Results of the univariate analysis for the vaccine exposure variables levels (IRR = incidence rate ratio)

	Exposure level	Total no. of flocks	No. of positive flocks	Incidence	IRR
				(new cases per 1000 farms)	(95%CI)
Categorization of the number of vets or retailers who had bought the vaccine	1 (0-1vet or retailer)	71 216	47	0.66	Referent
	2 (>1 vet or retailer)	7247	10	1.4	2.1 (1.1-4.1)
Categorization of the number of farms that had bought the vaccine	1 (0–4 farms)	49 849	25	0.50 (*)	Referent
	2 (5–10 farms)	20 821	17	0.82 (*)	1.6 (0.88-3.0)
	3 (>10 farms)	7793	15	1.9 (*)	3.8 (2.0-7.3)
Herd size (median for each province)		78 463	57		1.005
					(1.001-1.01)

Both exposure variables related to the vaccine [(1) categorization of the number of veterinarians or retailers who had purchased the vaccine and (2) categorization of the number of farms that had bought the vaccine] show a significant association with disease levels. The categorization of the number of farms that had bought the vaccine shows a statistically significant upward trend in risk. The median farm size for each province resulted as not having a health impact on the disease risk.
(*) Test for trend: χ2 = 2.92, P value = 0.003.

ISSN: 1297-9716