Methods

A bibliographic search was performed in the Medline and Embase databases for all studies published between January 2001 and December 2007. The search strategy was the same as employed in a previous review and included the following keywords: (“adverse drug reaction reporting system” or “drug therapy/adverse effects” or “pharmaceutical preparations/adverse effects”) AND (“child” or “child-preschool”) AND “prospective studies”.3 The following criteria were applied to the resulting studies to identify those which were relevant: the patients studied were not selected for particular conditions or specific drug exposures, prospective monitoring was used to identify ADRs, and sufficient information was reported to calculate the incidence of those ADRs.

Studies that evaluated adverse drug events (ADEs) were excluded. An adverse drug event is “any injury resulting from medical interventions related to a drug”10 and includes both ADRs in which no error occurred and complications that result from medication errors. A manual search of the bibliographies of the selected article was also carried out to identify additional pertinent studies.

Meta-analytic weighted average and 95% CIs of the incidence of ADRs were estimated using a random effect model to take into account the heterogeneity of the various studies.3 11

Statistical analysis was performed using SAS software v 9.1.

The websites of international drug regulatory agencies were searched to collect information on safety alerts concerning the paediatric population issued during 2001–2007. The websites of the following agencies were searched: Medwatch, United States (http://www.fda.gov/medwatch/index.html); Health Canada, Canada (http://www.hc-sc.gc.ca); European Medicines Agency (EMEA), Europe (http://www.emea.europa.eu); Agence Française de Securite Sanitaire des Produits de Sante (AFSSAPS), France (http://www.afssaps.fr/); Agenzia Italiana del Farmaco (AIFA), Italy (www.agenziafarmaco.it); Bundesinstitut für Arzneimittel und Medizinprodukte (BfarM), Germany (http://www.bfarm.de); Agencia Española de Medicamentos y Productos Sanitarios (AEMPS), Spain (http://www.agemed.es); Medicines and Healthcare products Regulatory Agency (MHRA), United Kingdom (http://www.mhra.gov.uk); Therapeutic Goods Administration, Australia (http://www.tga.gov.au); and New Zealand Medicines and Medical Devices Safety Authority (Medsafe) (www.medsafe.govt.nz).

Results

Review of the literature

The bibliographic search identified 57 papers, 46 of which were excluded because they were not pertinent, mainly because they concerned specific drug exposures (24 papers). One additional study was identified after the manual search.12 Three studies performed in the United States were subsequently excluded because their data involved ADEs and not ADRs and one because adults were also involved.13 14 15 16

Eight studies were therefore evaluated (table 1).12 17 18 19 20 21 22 23 Prospective paediatric drug surveillance studies were performed in six different countries, with two each in Germany and France. A majority of the reports concerned ADR incidence in hospitalised children (6/8), four dealt with ADRs in children leading to hospital admission, and two reported the incidence of ADRs in outpatient children. Two papers reported data regarding ADRs in hospitalised patients and leading to admission12 21 and one reported data from all three settings.17 The incidence of ADRs ranged from 1.5% to 19.9% among the studies involving hospitalised children.

The rate of hospitalisation due to ADRs ranged between 0.6% and 6%, while the incidence in outpatient children ranged between 0.7% and 11%. The meta-analytic estimated average, adjusted and weighted by sample size, was 10.9% (95% CI 4.8 to 17.0) for ADRs in hospitalised children, 1.8% (95% CI 0.4 to 3.2) for ADRs leading to hospital admission, and 1.0% (95% CI 0.3 to 1.7) for ADRs in outpatient children (table 2).

The severity of events was reported in five studies12 18 20 21 23: in two no severe ADRs were reported,20 23 while in the other three the percentage of severe ADRs ranged between 2% and 30%. Six studies reported the criteria for causality assessment,12 18 19 20 21 23 five of which used the WHO criteria.12 18 19 20 21 In these studies the ADRs with a cause assessed as definite/probable ranged between 56% and 91%.

Two studies evaluated the risk of ADRs linked to off-label drug use and found that the risk of ADRs was 3.6-fold higher in children and adolescents receiving off-label drugs in the hospital setting19 and 2-fold higher in the outpatient setting.23 The skin and gastrointestinal system were the organ systems most commonly involved and antibiotics were the drugs most commonly associated with ADRs (table 1). Only three studies reported the active substances most frequently associated with ADRs18 21 23: oxacillin,18 vancomycin21 and amoxi-clavulanic acid23.

Regulatory agency measures

In 2001–2007, the drug regulatory agencies listed above issued alerts aimed at health professionals and/or the public about the safety of medications in the paediatric population for a total of 28 drugs or drug classes (table 3). For 12 of these drugs, warnings were published in 2006–2007 alone. Most of the warnings were issued by the Food and Drug Administration (20 drugs), followed by the AIFA (11 drugs), the AFSSAPS and the MHRA (nine drugs each).

On the EMEA website only information concerning the reviews performed by the Committee for Medicinal Products for Human Use on selective serotonin reuptake inhibitors (SSRIs) and oseltamivir was found. On the website of the Spanish drug regulatory agency only warnings concerning SSRIs and aspirin were found, while on the websites of New Zealand and Australian agencies only warnings involving antidepressants were retrieved; these countries were therefore not taken into account in the analysis. A warning about antidepressants was provided in all the countries evaluated, while in five out of six countries warnings were issued for both intranasal desmopressin and ceftriaxone.

For eight drugs, alerts were issued only by the FDA. Warnings concerning immunostimulants and NSAIDs were issued only in France, cefaclor and domperidone in Italy, and halothane in the UK.

In most instances, the measure taken by the regulatory agencies consisted of a “dear health professional letter” and the modification of a drug label. For 10 drugs (35%), the measures were more drastic: pemoline, amphetamine and immunostimulants were withdrawn from the market (in the United States, Canada and France, respectively) and intranasal desmopressin’s indication for nocturnal enuresis was withdrawn in the US, France and Italy.

The use of promethazine and cough and cold medicines was contraindicated in children under 2 years of age in the United States, while that of nasal decongestants was contraindicated under 12 years of age in Italy. The use of aspirin was contraindicated under 16 years of age in the UK and in Italy due to the risk of Reye’s syndrome and that of metoclopramide was contraindicated in children under 16 years of age in Italy. A “black box” warning was added to antidepressants and atomoxetine in the USA.

A comparison between the list of drugs associated with ADRs in the prospective studies and the list of drugs for which a safety alert was issued was performed to evaluate the degree of overlap.

The prospective studies reported an ADR only for domperidone,23 metoclopramide,21 promethazine19 (CNS disorders) and aspirin (rash).12

Discussion

The number of prospective studies published in 2001–2007 seems to indicate an increased awareness of the importance of monitoring ADRs in the paediatric population. Eight studies were retrieved over the surveyed 8-year period versus 17 studies included in the meta-analysis by Impicciatore et al that covered the period between 1973 and 2000.3 The increased number of safety warnings issued by the drug regulatory agencies, in particular by the FDA, may be an index of greater attention paid to the issue of drug safety in children and adolescents.

Internationally, initiatives addressing drug safety have recently been set up, and this also indicates an increased prioritisation of the issue. Guidelines on pharmacovigilance in the paediatric population were adopted by the EMEA in 2006 and came into effect in January 2007 along with the European regulation on medicinal products for paediatric use. In 2007 the document Promoting safety of medicines for children was published by the WHO with the aim of pressing governments, industries and healthcare professionals to improve monitoring the safety of drug use in the paediatric population.24

ADR rates are similar to those reported in the meta-analysis by Impicciatore et al.3 This means that the incidence of ADRs has remained constant since 2001 and that hospitalised children have a 5-fold higher risk of developing an ADR compared to the outpatient paediatric population.

The finding concerning the increased risk of ADRs in children receiving unlicensed or off-label drug prescriptions confirms the results of previous studies25 26 and calls for more initiatives aimed at updating knowledge through clinical trials. The skin and gastrointestinal tract were the organ systems most commonly involved and antibiotics were the drugs most frequently associated with ADRs in children.

Taking into account the drugs most frequently associated with ADRs in prospective studies and those for which a safety alert was issued, the overlap between the two lists was poor.

This finding is consistent with the fact that the majority of the ADRs observed in the prospective studies were mild, extensively acknowledged and reported, and concerned drugs with a high prevalence of use in children. On the contrary, safety alerts almost exclusively concerned serious adverse effects, most of which were unknown or underestimated and concerned drugs with a low prescription prevalence.

Psychotropic drugs were the group of drugs for which the most warnings were issued and all the regulatory agencies issued warnings for them. Psychotropic drug use in children and adolescents has been a widely debated topic in the scientific community in the past few years, especially regarding SSRIs and their link to an increased risk of suicidality.27 28 The matter of antidepressant drug use in the paediatric population also relates to efficacy. In fact, the benefit of SSRIs seems to vary according to the disorder,29 and is particularly controversial in the treatment of major depressive disorder, for which the risk–benefit profile appears favourable only for fluoxetine.30 31 32 With the exception of a few drugs (sertraline and fluvoxamine for obsessive compulsive disorder, and fluoxetine for major depressive disorder), SSRIs are not licensed for children. The case of antidepressant use in children and adolescents therefore seems to highlight the fact that some therapies commonly used in the adult population are often applied to the paediatric population in an off-label/unlicensed manner.

It is interesting to note that some warnings concern drugs licensed for children, but with scant evidence of safety and effectiveness in the paediatric population. The alerts concerning the use of cough and cold medicines, often sold as over-the-counter drugs, are noted. More than 1500 children aged <2 years were treated in US emergency departments in 2004–2005 for adverse events associated with cough and cold medications, including overdoses. Three cases of infant death were also reported and all three had high levels of pseudoephedrine in postmortem blood samples.33 Moreover, a review by the FDA identified 123 deaths related to the use of such products in children under 6 years of age during the past few decades.34 These risks are unacceptable considering that the efficacy of these drugs has not been proven and remains controversial.35 36 37 38

Another example is represented by anti-emetic drugs (metoclopramide and domperidone), for which cases of dystonia were reported in the prospective studies and warnings were issued in Italy and France. In 2007, the AIFA Paediatric Working Group reviewed data concerning the drug therapies for emesis in childhood and concluded that evidence on the efficacy of metoclopramide and domperidone in the treatment of acute emesis (gastroenteritis and gastro-oesophageal reflux) is scant and controversial while the risk of drug toxicity is widely documented.39 Some alerts were issued only by one drug regulatory agency. This could in part be due to greater attention paid to the problem of drug safety by some agencies (eg, FDA) and in part to the inter-country differences in the licensing and marketing of drugs and in the practitioners’ prescribing habits.

In this context, it is not surprising that an alert concerning cefaclor was published only in Italy, since in this country, antibiotics and, in particular, cephalosporins are widely prescribed drugs in children. Cefaclor is less frequently prescribed in other European countries such as the UK or the Netherlands.40 41 Another case is the bacterial lysate used as an immunostimulant in patients with recurrent infections. In 2005, AFSSAPS decided to withdraw such products from the market since evidence of their efficacy was limited and cases of severe allergic reactions had been reported. It is not by chance that France was the only country to issue an alert, since these drugs are not licensed in the USA or UK. However, these medications are still marketed in other countries (eg, Italy) and are frequently prescribed to children. Finally, the described differences between warnings and the published evidence concerning drug safety may be also related to a mixture of science, surveillance and other national drug policy factors that affect the decisions taken at the regulatory level.

Conclusions

Efforts are needed at different levels (governments, drug regulatory agencies, pharmaceutical industries, health care professionals and parents) to ensure the safe use of medications in the paediatric population, not only in individual countries but worldwide, through common and collaborative initiatives. Prospective multinational studies should be planned and supported and common regulatory actions (eg, safety alerts) should be taken with the aim of improving the rational use of drugs in children.