Risk factors for secondary transmission of Shigella infection within households: implications for current prevention policy

Background: Internationally, guidelines to prevent secondary transmission of Shigella infection vary widely. Cases, their contacts with diarrhoea, and those in certain occupational groups are frequently excluded from work, school, or daycare. In the Netherlands, all contacts attending pre-school (age 0-3) and junior classes in primary school (age 4-5), irrespective of symptoms, are also excluded pending microbiological clearance. We identified risk factors for secondary Shigella infection (SSI) within households and evaluated infection control policy in this regard. Methods: This retrospective cohort study of households where a laboratory confirmed Shigella case was reported in Amsterdam (2002-2009) included all households at high risk for SSI (i.e. any household member under 16 years). Cases were classified as primary, co-primary or SSIs. Using univariable and multivariable binomial regression with clustered robust standard errors to account for household clustering, we examined case and contact factors (Shigella serotype, ethnicity, age, sex, household size, symptoms) associated with SSI in contacts within households. Results: SSI occurred in 25/337 contacts (7.4%):20% were asymptomatic, 68% were female, and median age was 14 years (IQR: 4-38). In a multivariable model adjusted for case and household factors, only diarrhoea in contacts was associated with SSI (IRR 8.0, 95% CI:2.7-23.8). In a second model, factors predictive of SSI in contacts were the age of case (0-3 years (IRRcase >= 6 years:2.5, 95% CI:1.1-5.5) and 4-5 years (IRRcase >= 6 years: 2.2, 95% CI:1.1-4.3)) and household size (>6 persons (IRR2-4 persons 3.4, 95% CI:1.2-9.5)). Conclusions: To identify symptomatic and asymptomatic SSI, faecal screening should be targeted at all household contacts of preschool cases (0-3 years) and cases attending junior class in primary school (4-5 years) and any household contact with diarrhoea. If screening was limited to these groups, only one asymptomatic adult carrier would have been missed, and potential exclusion of 70 asymptomatic contacts <6 years old from school or daycare, who were contacts of cases of all ages, could have been avoided