Search in the Abstract Database

* = Presenting author

P741 Fine-scale geographic distribution and ecological studies of Crohn’s disease in France (2007–2014)

M. Genin*1, M. Fumery2, F. Occelli3, G. Savoye4, B. Pariente5, L. Dauchet6, C. Vignal7, M. Body-Malapel7, J. Giovannelli8, H. Sarter9, C. Gower-Rousseau8, G. Ficheur10

1University of Lille, EA2694 - Santé publique : épidémiologie et qualité des soins, Lille, France, 2Amiens University Hospital, Gastroenterology Unit, Amiens, France, 3University of Lille, EA 4483 - Impact de l’environnement chimique sur la santé humaine, Lille, France, 4Rouen University Hospital, Gastroenterology Unit, Rouen, France, 5Lille University Hospital, Gastroenterology Unit, Lille, France, 6Lille University hospital, INSERM, UMR1167, Lille, France, 7University of Lille, UMR 995, Lille, France, 8Lille University Hospital, UMR 995, Lille, France, 9Lille university Hospital, EA 2694 - Santé publique : épidémiologie et qualité des soins, Lille, France, 10Lille University Hospital, EA 2694 - Santé publique : épidémiologie et qualité des soins, Lille, France

Background

A significant geographical variation in the incidence of Crohn's disease (CD) suggests the role of environmental factors in its pathogenesis. The objectives of this work were (i) to study the spatial distribution of CD cases in France from the database of the Programme de Médicalisation des Systèmes d'Information (PMSI), (ii) to study the factors associated with spatial heterogeneity and (iii) to identify clusters of high-prevalence.

Methods

Patients with a CD diagnostic code were searched in the PMSI database between 2007 and 2014. To study the spatial distribution of prevalence for this period, a fine-scale spatial unit (5610 units at the national level) was used. The results were expressed as standardised prevalence ratio (SPR). An ecological regression measured the association between the risk of CD and spatial unit characteristics, such as access to the nearest reference centre, urbanisation and rurality, deprivation index (FDep index), latitude, and distance from polluting facilities. Elliptical spatial scan statistics were used to search high-prevalence clusters.

Results

A total of 129089 patients with CD have been identified, with a national prevalence in 2014 of 203/100000 inhabitants. Significant spatial heterogeneity has been found (p ≤ 0.0001) (Figure 1). Ecological regression revealed a significant association between the risk of CD and the highest deprivation index (Relative Risk (RR) = 1.05[1.02–1.08]) and urbanisation (RR=1.08[1.05–1.15]). The spatial analysis detected 16 clusters with a CD relative risk ranging from 1.40[1.31–1.50] to 1.90[1.65–2.19].

Geographical distribution of age-and-gender Standardised Prevalence Ratios of Crohn’s disease in France, 2007–2014.

Conclusion

The French geographical distribution of CD is not uniformly distributed. Sixteen clusters with high-risk of CD have been identified. The poorest populations and industrialised areas were associated with a higher risk of developing CD.