From 1998 to 2007 a total of 2437 individual ADR reports containing information about 4500 ADRs were reported for children. Of these, 210 reports corresponding to 429 ADRs were submitted for psychotropic medicines. Figure 1 shows the annual distribution of the reported ADRs. There were wide fluctuations in the number of ADRs reported annually, with an increase in number from 2003 to 2005 followed by a decrease in 2006 and 2007.

Table 1 shows the distribution of reported ADRs by therapeutic group and medicine, age of patient and seriousness. Almost one fifth of ADRs were reported for children below 2 years and one half of all ADRs were reported for adolescents (from 11 to 17 years of age), and 45% of these were serious. Totally, 59% of all ADRs were reported for boys. More than one half of all ADRs were classified as 'serious'. Table 2 displays characteristics of ADRs reported for children below two years. Two deaths were reported for citalopram and fluoxetin due to chorioamnionitis and persistent foetal circulation, respectively. ADRs among children up to 2 years of age encompassed a wide range of reactions, e.g. convulsion, feeding disorder, neonatal priapism, apnoea and ventricular septal defects. Seven ADRs: drug exposure during pregnancy, neonatal respiratory depression, apnoea and pallor were reported as "maternal drugs affecting foetus". However, the remaining ADRs were probably also caused by mothers' use of psychotropic medicines during pregnancy, as the indications for use were reported as depression, anxiety, panic disorder and schizophrenia. The share of serious psychotropic ADRs was higher than the share of serious ADRs in Danish children in general (40%) [¹¹]. In the general Danish paediatric population, half of all ADRs were reported in children from birth up to 2 years of age, but for psychotropic medicines more than half of all ADRs were reported for adolescents which reflect a more extensive use of psychotropic medicine in 11 to 17- year-olds [⁹]. The majority of serious ADRs were reported in infants in contrast to reports submitted to Health Canada where 60% of all ADRs for psychotropic medicines were reported in 13 to 19-year-olds, and only 12% in infants [¹²]. A number of ADRs were reported for Danish children below 2 years, probably due to the mother's intake of psychotropic medicine, primarily antidepressants and antipsychotics, during pregnancy. Serious ADRs such as 'neonatal withdrawal syndrome', 'ventricular septal defects' and 'premature labour' were reported. The risks of malformations as well as preterm delivery of babies due to use of antidepressants during pregnancy have been reported previously in the literature and is supported by our results [^13-16].

The largest share of ADRs (42%) was reported for psychostimulants (ATC group N06B), followed by 31% for antidepressants (ATC group N06A) and 24% for antipsychotics (ATC group N05A). More than one half of the ADRs reported for antipsychotics were caused by the drugs ziprasidone, olanzapine and risperidone. Although only 2.5% of ADRs were reported for anxiolytics and sedatives (ATC group N05B and N05C), predominantly in infants, all these ADRs were serious. Two-thirds of the ADRs reported for antidepressants (ATC group N06A) were reported for infants and adolescents and exclusively for the medicines sertraline, citalopram and fluoxetine and seventy-five percent of these were serious. For psychostimulants (ATC group N06B) 50% of ADRs were serious and reported for children from six to nine years of age, and 40% of the reports were associated with methylphenidate and atomoxetine. With one exception, all 70 ADRs reported for children less than one year of age were serious.

Table 3 shows the distribution of reported ADRs by system organ class (SOC). The largest shares of ADRs were reported for the SOCs 'psychiatric disorders' (20% of total), 'nervous system disorders' (20% of total) and 'general disorders and administration site conditions' (12% of total). Less than 1% of the total number of reports concerned the following SOCs: 'ear and labyrinth disorders', 'endocrine disorders', 'hepatobiliary disorders', 'immune system disorders', 'pregnancy, puerperium and perinatal conditions' and 'surgical and medical procedures'. The largest shares of serious ADRs, about 25% of all were reported for the SOCs: 'psychiatric disorders' and 'nervous system disorders'. The distribution between serious and non-serious ADRs within SOCs varied. More than seventy-five percent of ADRs reported from the SOCs: 'nervous system disorders', 'musculoskeletal and connective tissue disorders' and 'congenital, familial and genetic disorders'were serious.

There are huge gaps in the evidence on the safety of medicines in children as only few medicines prescribed for children are tested in clinical trials and licensed for use in this population. In addition, information about serious and long-term ADRs is sparse due to the limitations embedded in the design of randomised, controlled clinical trials which are used primarily to test hypotheses about efficacy rather than safety and children are usually excluded from clinical trials of medicines for ethical reasons [¹⁷]. Therefore, it is very important to systematically analyse and evaluate data reported to the spontaneous reporting programmes as these reports are the major source for new information about possibly serious and previously unknown ADRs [¹⁸]. Several of the reported ADRs were birth defects, an area where we have very limited knowledge.

*Indication for use not listed, but recorded as "Maternal drugs affecting foetus"; NA: not available