Background Occult rib fractures in infancy have the highest positive predictive value for physical abuse (PA) of all fractures. However, they often pose a clinical diagnostic dilemma. We describe the differential diagnosis of rib fractures in infancy and propose a diagnostic algorithm.

Methods Children under the age of 24 months with rib fractures were identified from radiological databases in three regional centres from 1998 to 2007. Cases were sorted into diagnostic groups and clinical features were compared and analysed using SPSS 18 for Mac OSX.

Results 52 children with rib fractures were identified. 17 had confirmed PA (mean age 4.5 months, range 1–15), 3 suffered witnessed trauma (mean age 16.3 months, range 5–24), 11 were post-surgical (mean age 2.7 months, range 0–5) and 21 were unknown cases (mean age 3.16 months, range 1–7) including 17 with presumed metabolic disease of prematurity, 3 children with no risk factors and 1 fracture presumed to be secondary to cardio-pulmonary resuscitation. When the PA and the unknown group were compared, there was no statistical difference in the number (p=0.131), location (p=0.073–0.525) or distribution of fractures and the NAI group were not more likely to suffer bilateral fractures. Furthermore, significantly more children in the unknown group had a gestational age <30 weeks (17/21) and a birth weight <1.25 kg (18/21) (p<0.001). The unknown group were also more likely to be on frusemide (9/21) and have chronic lung disease (5/21) or prolonged total parenteral nutrition (TPN) > 28 days (8/21). Three children had no predisposing factors for metabolic bone disease and PA was excluded after child protection assessment. All had single rib fractures noted on pre-operative CXRs (1 was on frusemide).

Conclusions The number, location, distribution and age of infants with rib of fractures are not discriminatory features for PA. A comprehensive assessment of an infant with occult rib fractures should include a child protection assessment and assessment for gestational age <30 weeks, low birth weight, chronic lung disease, frusemide, prolonged TPN and history of CPR. (Risk factors and clinical indicators for OI should be considered but were not illustrated in this case series).