Objective To determine the frequency of non-cutaneous mimics identified in a large, multicentre cohort of children evaluated for physical abuse.

Methods Prospectively planned, secondary analysis of 2890 physical abuse consultations from the Examining Siblings To Recognize Abuse (ExSTRA) research network. Data for each enrolled subject were entered at the child abuse physician's diagnostic disposition. Physicians prospectively documented whether or not a ‘mimic’ was identified and the perceived likelihood of abuse. Mimics were divided into 3 categories: (1) strictly cutaneous mimics, (2) strictly non-cutaneous mimics and (3) cutaneous and non-cutaneous mimics. Perceived likelihood of abuse was described for each child on a 7-point scale (7=definite abuse).

Results Among 2890 children who were evaluated for physical abuse, 137 (4.7%) had mimics identified; 81 mimics (59.1% of mimics and 2.8% of the whole cohort) included non-cutaneous components. Six subjects (7.4%) were assigned a high level of abuse concern and 17 (20.1%) an intermediate level despite the identification of a mimic. Among the identified mimics, 28% were classified as metabolic bone disease, 20% haematologic/vascular, 16% infectious, 10% skeletal dysplasia, 9% neurologic, 5% oncologic, 2% gastrointestinal and 10% other. Osteomalacia/osteoporosis was the most common non-cutaneous mimic followed by vitamin D deficiency.

Conclusions A wide variety of mimics exist affecting most disease categories. Paediatric care providers need to be familiar with these conditions to avoid pitfalls in the diagnosis of physical abuse. Identification of a mimic does not exclude concurrent abuse.