A Score to Predict ADHD in Preschoolers Based on Neonatal Data

Prematurity is an established risk factor for the development of Attention Deficit/Hyperactivity Disorder (ADHD). The purpose of our study was to evaluate the association of ADHD symptoms with perinatal factors and to develop a risk score.

We performed an evaluation protocol of ADHD diagnosis in a cohort of 96 premature children assisted in the Neonatology Intensive Care Unit of the Hospital de Clínicas de Porto Alegre, born between January, 2010 and December, 2012 when they were at 4 to 6 years old. ADHD was diagnosed using a semi structure interview (KSDAS-PL). Sociodemographic and perinatal data were collected. Patients who developed and who did not develop ADHD were compared according to perinatal factors selected from the literature and a previous metaanalyses conducted by our group. Maternal antecedents: family sociodemographic classification, family history of ADHD, ASRS score of parents, alcohol use, tobacco or any illicit drug during pregnancy, maternal postpartum depression. Gestational data: Group B streptococcus colonization, gestational diabetes mellitus or diabetes mellitus, chronic hypertension, urinary tract infection, preeclampsia. Newborn data: retinopathy of prematurity, periintraventricular hemorrhage, any neurological abnormality, hospital discharge with O2, Apgar at 1 'and 5', apnea, days of life in which birth weight was regained, bronchopulmonary dysplasia, ultrasound at hospital discharge, hypoxic-ischemic encephalopathy, neurological examination at hospital discharge, pulmonary hemorrhage, hypoglycemia, final gestational age, periventricular leukomalacia, percentile <3, head circumference, birth weight, early or hospital sepsis, gender, SNAPPE II - Acute Neonatal Physiology with Perinatal Extension-II, TORCHS infections, total parenteral nutrition time, non-invasive ventilation time, cephalic perimeter at hospital discharge, oxygen therapy time, mechanical ventilation time, neonatal seizures and necrotizing enterocolitis and chorioamnionitis.

Taking into account the impact of each variable in relation to the emergence of ADHD, the following variables remained in the model to build a continuous value score to predict ADHD (accepting a flexible p value – 0.2): head circumference at hospital discharge, bronchopulmonary dysplasia, neonatal seizures, necrotizing enterocolitis, and chorioamnionitis. The most significant variable was necrotizing enterocolitis. For the establishment of sensitivity / specificity, the ROC curve including these 5 variables found a satisfactory performance, with an area under the ROC of 0.707.

The study was approved by the Hospital de Clínicas Ethics Committee. All individuals and parents fulfilled a standard informed consent term.