[Home falls in infants before walking acquisition]

Minor head trauma is frequent among infants and leads to numerous visits to emergency departments for neurological assessment to evaluate the value of cerebral CT scan with the risk for traumatic brain injuries (TBI).

To analyze the epidemiological characteristics of nonwalking infants admitted after falling at home and to analyze associated factors for skull fractures and TBI.

Between January 2007 and December 2011, all children aged 9 months or younger and admitted after a home fall to the pediatric emergency unit of a tertiary children's hospital were included. The data collected were age, sex, weight and height, body mass index; geographic origin, referral or direct admission, mode of transportation; month, day and time of admission; causes of the fall, alleged fall height, presence of an eyewitness, type of landing surface; Glasgow Coma Scale (GCS) score, application of the head trauma protocol, location and type of injuries, cerebral CT scan results, length of hospital stay, progression, and neglect or abuse situations.

DESCRIPTIVE ANALYSIS: within the study period, 1910 infants were included. Fifty-four percent of children were aged less than 6 months with a slight male prevalence (52%). Falls from parental bed and infant carriers accounted for the most frequent fall circumstances. GCS score on admission was equal to 14 or 15 in 99% of cases. A cerebral CT scan was performed in 34% of children and detected 104 skull fractures and 55 TBI. Infants aged less than 1 month had the highest rate of TBI (8.5%). Eleven percent of patients were hospitalized. A situation of abuse was identified in 51 infants (3%). UNIVARIATE ANALYSIS: Male children and infants aged less than 3 months had a higher risk of skull fractures (P = 0.03 and P = 0.0003, respectively). In the TBI group, children were younger (3.8 ± 2.6 months versus 5.4 ± 2.5 months, P < 0.0001), fell from a higher height (90.2 ± 29.5 cm versus 70.9 ± 28.7 cm, P < 0.0001), were more often admitted on a weekend or day off, and had more skull fractures (54% versus 6%, P < 0.001). MULTIVARIATE ANALYSIS: all variables showing P < 0.2 in the univariate analysis were entered into the model. In the final model, three variables continued to be associated with a risk of TBI: being referred by a physician (OR 4.6 [2.2-9.6], P < 0.0001), being younger than 3 months old (OR 3.1 [1.7-5.7], P = 0.0002), falling from a height greater than 90 cm (OR 3.1 [1.7-5.6], P = 0.0002).

Before walking acquisition, children are particularly vulnerable and have the highest rate of TBI after a vertical fall. In this age group, the rate of abuse is also higher. Given this double risk, numerous cerebral CT scans are performed (35-40% of the target population). This protocol, however, leads to a low proportion of detected TBI (<10%) compared to the high number of CT scans and an additional risk of irradiation.

As no validated predictive score exists and pending the contribution of the S-100B protein assay, the identification of infants at high risk for TBI and justifying neuroimaging is based on the search for predisposing factors and circumstances.