Objective To determine the significance of bruises in cases of infant death to ascertain if their presence was associated with other injuries or homicide.

Design Retrospective cohort study.

Setting Forensic Science South Australia, Adelaide, South Australia, Australia.

Patients 135 infants aged between 7 and 364 days autopsied over a 9-year period from June 1999 to May 2008.

Outcome measures Coronial forensic autopsy reports.

Methods Examination of autopsy reports to determine how many infants had bruising and whether bruising was associated with other injuries or lesions of concern. Cause and manner of death were also recorded.

Results There were 83 boys and 52 girls. Twenty-one infants (15.6%) had one or more bruises, and 114 had no bruises. In the group of 21 infants with bruises, 17/21 had other injuries/lesions (81%), with 5/21 homicides (24%). Of the 114 non-bruised infants, only 9 (8%) had other injuries/lesions with only 4 homicides (3.5%). The incidence of other injuries/lesions was significantly higher in bruised infants compared with the non-bruised group (p<0.001), as was the occurrence of homicide (p=0.003). Bruises were found in 15.6% of infants presenting to coronial autopsy, with a sensitivity of 65% and a specificity of 96% as markers for other injuries (positive predictive value 0.81; negative predictive value 0.92). At <6 months of age, this increased to a sensitivity of 71%, with a similar specificity of 95%.

Conclusions Bruising is a significant marker for other injuries in infants presenting to coronial autopsy. Homicide occurs more commonly in this group.