%0 Journal Article
%9 ACL : Articles dans des revues à comité de lecture répertoriées dans les bases internationales
%A Gérardin, P.
%A Guernier, V.
%A Perrau, J.
%A Fianu, A.
%A Le Roux, K.
%A Grivard, P.
%A Michault, A.
%A Lamballerie de, Xavier
%A Flahault, A.
%A Favier, F.
%T Estimating Chikungunya prevalence in La Réunion Island outbreak by serosurveys : two methods for two critical times of the epidemic
%D 2008
%L fdi:010048276
%J Bmc Infectious Diseases
%M ISI:000259221500001
%P 99
%R 10.1186/1471-2334-8-99
%U http://www.documentation.ird.fr/hor/fdi:010048276
%> http://www.documentation.ird.fr/intranet/publi/2009/compl2/010048276.pdf
%V 8
%W Horizon (IRD)
%X Background: Chikungunya virus (CHIKV) caused a major two-wave seventeen-month-long outbreak in La Reunion Island in 2005-2006. The aim of this study was to refine clinical estimates provided by a regional surveillance-system using a two-stage serological assessment as gold standard. Methods: Two serosurveys were implemented: first, a rapid survey using stored sera of pregnant women, in order to assess the attack rate at the epidemic upsurge (s1, February 2006; n = 888); second, a population-based survey among a random sample of the community, to assess the herd immunity in the post-epidemic era (s2, October 2006; n = 2442). Sera were screened for anti-CHIKV specific antibodies (IgM and IgG in s1, IgG only in s2) using enzyme-linked immunosorbent assays. Seroprevalence rates were compared to clinical estimates of attack rates. Results: In s1, 18.2% of the pregnant women were tested positive for CHIKV specific antibodies (13.8% for both IgM and IgG, 4.3% for IgM, 0.1% for IgG only) which provided a congruent estimate with the 16.5% attack rate calculated from the surveillance-system. In s2, the seroprevalence in community was estimated to 38.2% (95% CI, 35.9 to 40.6%). Extrapolations of seroprevalence rates led to estimate, at 143,000 and at 300,000 (95% CI, 283,000 to 320,000), the number of people infected in s1 and in s2, respectively. In comparison, the surveillance-system estimated at 130,000 and 266,000 the number of people infected for the same periods. Conclusion: A rapid serosurvey in pregnant women can be helpful to assess the attack rate when large seroprevalence studies cannot be done. On the other hand, a population-based serosurvey is useful to refine the estimate when clinical diagnosis underestimates it. Our findings give valuable insights to assess the herd immunity along the course of epidemics.