Is steroid use in premature babies linked to ADHD?

Attention Deficit Hyperactivity Disorder (ADHD) is the most common pervasive developmental disorder amongst children of school age and it is therefore not surprising that researchers will look to determine factors that may be linked to or even cause the disorder. The findings of a recently published study with a Finnish sample have been interpreted by the media as showing a link between steroid use in premature infants and the development of ADHD. How credible is this link?

Before we can evaluate this link, it is important to address what ADHD actually is. The disorder is characterised by persistent symptoms of inattention, hyperactivity and impulsivity that impact negatively on many aspects of functioning and development including family, social and academic areas of a child’s life. ADHD symptoms will be evident in at least two settings, typically school and home. ADHD is common; in the UK estimates range between 3-9%. ADHD often continues into adulthood with an estimated 2-4% of adults still presenting with significant symptoms. The UK charity ADDISS provide information, support and resources in relation to ADHD.

What did the researchers of this study do?

This study which examined the link between steroid use in premature infants and the later development of ADHD is excellently described in an appraisal of the study reported on the NHS Choices website Behind the Headlines. A full description of the study in lay terms can be accessed here.

Caution when interpreting the results

As pointed out in the Behind the Headlines appraisal of this study, the small sample size of the study (37 children followed up at age 8 and 29 at age 16) means that the findings need to be considered exploratory. The research has many strengths, though, including careful matching of children and control of a range of confounding factors (factors other than use of steroids that could instead explain the results).

A significant limitation of the study, however, concerns the measurement of ‘ADHD’. As described above, a diagnosis of ADHD requires verification of symptoms in 2 settings, typically by parents and teachers. While the study includes both of these, this data is in fact collected at the different age points assessed in the study (teacher data is reported at age 8 and parent data at age 16). To determine if steroid use is linked to an actual diagnosis of ADHD, the data should have been collected by both parents and teachers at the same time point. The link between prematurity and ADHD symptoms at age 16 is not significant in any case, meaning that a link is being made between steroid use and teacher reports of ‘ADHD related symptoms’ only. I use the term ‘ADHD related symptoms’ because the data at age 8 was measured using a child mental health screener that is not specific to ADHD. Furthermore, the authors do not confirm at any point whether the children indeed received a diagnosis of ADHD at any point within child psychiatric/pediatric services.

How did the media handle coverage of this study?

Both the Telegraph and BBC, who covered the study findings, were careful to set the study in context, such as noting the researcher’s comments that the benefits of the steroids for premature infants outweigh the potential harms. Both avoided making a direct causal relationship between steroid use and ADHD—the BBC stated that steroids ‘may raise risk of ADHD’ and the Telegraph similarly had the cautious headline ‘steroid injections for premature babies could raise ADHD risk’.

In previous posts, we at Research the Headlines have noted the use of inappropriate headlines that sensationalise study findings, so it is very positive here to see the careful wording used to characterise the findings. There are a couple of discrepancies in the media coverage, however. The Telegraph stated that those whose mothers had been treated performed worse on average in general mental health tests at ages eight and 16. However, only the findings at age 8 were actually significant.  The BBC stated that when the children were followed to the age of eight, there was a higher incidence of attention deficit hyperactivity disorder. The terminology used here, in particular the use of ‘disorder’ rather than ‘symptoms’ suggests the findings were linked to a diagnosis of the disorder which they were not.

Khalife N, Glover V, Taanila A, Ebeling H, Järvelin M-R, et al. (2013) Prenatal Glucocorticoid Treatment and Later Mental Health in Children and Adolescents. PLoS ONE 8(11): e81394. doi:10.1371/journal.pone.0081394