Mini Oral Australian Epidemiology Association ASM 2018

A total population sibling linkage study of infection-related hospitalisation (#14)

Jessica E Miller 1 , Kim Carter 2 , Nicholas de Klerk 2 , David Burgner 1 3 4
  1. Murdoch Children's Research Institute, Parkville, VICTORIA, Australia
  2. Telethon Kids Institute, Subiaco, Western Australia, Australia
  3. Paediatrics, University of Melbourne, Parkville, Victoria, Australia
  4. Paediatrics, Monach University, Clayton, Victoria, Australia

Background

All children are exposed to pathogens, but only a minority develop severe infection. This differential susceptibility partly reflects shared heritable and environmental factors.

Aim

We aimed to estimate the sibling risk ratios of infection-related hospitalisations (IRHs) in children.

Methods

We analysed all live-born births from population-based Western Australian registries 1980-2014. Hospitalisations within families were linked with the Family Connections Project. We estimated sibling risk ratios; the ratio of risk of IRHs (overall and specific, defined by discharge ICD coding) in children (the proband) whose siblings had IRHs to the risk in probands whose siblings did not have IRHs. Adjusted RRs were estimated by Cox regression models. Children were followed from when they became a sibling until an IRH diagnosis (up to three diagnoses), death, 18th birthday, or end-2014.

Results

Among 536,117 probands, 155,885 had at least one IRH. 141,167 probands had a sibling with a previous IRH. The median time between sibling and proband IRHs was 1.4 years. Overall IRH risk increased with sibling IRHs: 1 sibling IRH (adjusted RR: 1.41, 95%CI: 1.39-1.43); 2 IRHs (1.65, 1.61-1.69) or 3+ IRHs (1.83, 1.77-1.90). Varying infection-specific effects were observed.

Conclusions

Having a sibling with a previous IRH was associated in a dose-response manner with a higher risk of IRH in the proband. Shared heritable and environmental factors are likely mechanisms. Sibling analyses measure associations from exposures at different ages, providing robust estimates of shared heritable determinants of infection that do not reflect seasonal variation in pathogen epidemiology and virulence.