Studying use and risks of medicines in children: a European approach

In recent years there has been increasing awareness on the gap in knowledge on the safety and efficacy of medicines used in children. The work as described in this thesis is based on large observational studies across Europe and aimed to study the use and risks of medicines in children and adolescents. First we studied prescriptions patterns of several classes of medicines in different European countries, second we assessed which adverse drug reactions are commonly reported, third we studied the methodological aspects of performing safety signal detection within paediatrics and last we studied the effect of several safety warning on the prescriptions of prokinetic drugs. Chapter 1 gives a general introduction to the different topics that were studied in this thesis and provides an overview of the data sources used. The chapter concludes with the outline of the thesis. Drug utilisation Chapter 2 describes the extent of primary care antibiotic prescribing in the Netherlands, the United Kingdom (UK) and Italy during a ≥10 years study period. We included almost 3 million children aged 0-<18 years, contributing over 12 million person years (PY) of follow up over a period of 10 to 16 years. The antibiotic prevalence rates amongst the studied countries varied with the Netherlands having the lowest prevalence (18.0 users/100 PY) and a two to three-fold higher prevalence rate for the UK (36.2 users/100 PY) and Italy (52.0 users/100 PY). Amoxicillin with or without clavulanic acid was the most common prescribed antibiotic. Strikingly, in Italy, cephalosporins, especially third-generation cephalosporins were commonly used – a phenomenon which was not observed in the other countries. The high consumption of cephalosporins in primary care in Italy is worrying, since the use of these broad-spectrum antibiotics is associated with higher rates of antimicrobial resistance. The results of this study underline the need for development of internationally agreed paediatric specific quality indicators for primary care prescribing of antibiotics. Chapter 3 describes the prescription pattern of oseltamivir, an antiviral drug, to children in primary care in the UK, Italy and the Netherlands during the influenza A(H1N1)pdm09 pandemic. Prescription patterns were compared with nationally reported rates for influenza like illness (ILI). The prevalence of oseltamivir prescriptions, stratified by calendar month, showed a biphasic pattern in all countries: A first peak was present in July 2009 and a second peak in October/November 2009. The ILI rates showed a similar pattern, with the highest rates in mid-July (155.3/100,000) in the UK, whereas the influenza A(H1N1)pdm09 epidemic reached its peak in both the Netherlands (189.4/100,000) and Italy (1,253.4/100,000) in the second week of November. Prevalence of oseltamivir prescriptions were highest for children aged 1 to 5 years (Netherlands: 2.6/1,000PM in November; UK: 10.7/1,000PM in July). The prevalence of use in Italy was limited (max. 0.2 users/1,000PM) while the ILI rates in Italy were highest. The marked variation in prescribing seems to be driven by public health policy. A European or worldwide approach or policy on recommendations on the use of antiviral drugs in future pandemics is warranted. In Chapter 4 the extent of prescribing proton pump inhibitors (PPIs) and histamine-2-receptor antagonists (H2RAs), two classes of gastric acid suppressant medicines, in children and adolescents in Spain, the Netherlands and Italy was studied. H2RAs were mainly prescribed in children up to age of 1 year in the Netherlands and Italy. PPIs were mainly prescribed in children of 12 years and older in Spain and the Netherlands (no data available in Italy). Within these two classes of gastric acid suppressant medicines, the share of PPI prescriptions increased significantly between 2001 and 2008. There is increasing evidence that long term use of PPIs in adults is associated with adverse events like bone fractures, community-acquired pneumonia and clostridium infections. With the increasing use of PPIs in children and adolescent, studies to the safety of long term use are warranted. Adverse drug reactions The results of the study on adverse drug reaction-related hospital admissions in children in the Netherlands are presented in Chapter 5. Adverse drug reactions (ADRs) accounted for 0.75% of all paediatric hospital admissions in the Netherlands in the period 2000 to 2005, with an incidence of 24/100,000.children in the population. Risks for ADR-related hospital admission were highest in children under the age of 2 years (incidence 169/100,000 children). These young children were mainly admitted for ADRs due to maternal exposure through placenta or breast milk. This underlines that paediatric safety is also extended to use of medicines in pregnancy. In Chapter 6 the safety experience and adverse events from 25 clinical studies and numerous analyses of spontaneous ADR reports for several pandemic H1N1 vaccines in children and adolescents is reviewed. At time of the influenza A(H1N1)pdm09 pandemic, these vaccines were licenced using fast track procedures, with relatively limited data on the safety in children and adolescents. However, the diversity in methods and data presentation in clinical study publications and publications of spontaneous reports hampered the analysis of safety of the different vaccines. As a result, relatively little has been learned on the comparative safety of these pandemic H1N1 vaccines - particularly in children. It should be a collective effort to give added value to the enormous work going into the individual studies by adhering to available guidelines for the collection, analysis, and presentation of vaccine safety data in clinical studies and to guidance for the clinical investigation of medicinal products in the paediatric population. Importantly the pandemic has brought us the beginning of an infrastructure for collaborative vaccine safety studies in the EU, US and globally. Chapter 7 described the paediatric individual case safety reports (ICSRs) reported to the Adverse Event Reporting System (AERS), as maintained by the US Food and Drug Administration (FDA). A total of 106,122 paediatric ICSRs (55% boys) (58% US) with a median of 1 medicine [range 0-157] and 3 events [1-94] per ICSR were described. Most commonly reported classes of medicines by decreasing frequency were ‘neurological’ (58%), ‘antineoplastic’ (32%) and ‘anti-infectives’ (25%). Most commonly reported system organ classes were ‘general’ (13%), ‘nervous system’ (12%) and ‘psychiatric’ (11%). We explored of these spontaneous reports are also suitable for studying long term effect of medicines. Unfortunately, duration of use could only be calculated for 19.7% of the reported medicines. Of these 14.5% concerned medicines being used long-term (>6 months). Systemical hormonal preparations, alimentary medicines and antineoplastic/immunomodulating agents were prominently reported after long-term treatment, while anti-infective medicines, musculoskeletal system drugs and sensory organ drugs were reported mostly with short term use. This is in line with what is expected based on indications and from knowledge on use of medicines. Large compilations of paediatric ICSRs might be a suitable additional source to generate signals on delayed events and new onset chronic events. Knowledge on the distribution of the classes of medicines and events within AERS is a key first step in developing paediatric specific methods for drug safety surveillance. Methods in paediatric safety signal detection In Chapter 8 we studied the effect of a large proportion of vaccine-related ICSRs in a spontaneous reporting database on the sensitivity of paediatric safety signal detection methods. We used VigiBase; vaccines made up 54% of all paediatric ICSRs reported in the period between 2000 and 2006. We compared the number of detected signals of disproportional reporting (SDRs) identified using either all paediatric reports available or using subgroups of vaccine-ADR pairs and non-vaccine-ADR. For ADRs reported less frequent for medicines (non-vaccines) than for vaccines, subgroup analysis using non-vaccines only led to an increase of the sensitivity of the signal detection process. If the ADR was reported more frequent for medicines than for vaccines, routine application of subgroup analysis can be harmful since then subgroup analysis led to a decreased sensitivity. Subgroup analysis can in this group be informative in specific subgroup analyses for studying the magnitude of class effects for the individual drugs in the class. Chapter 9 explored the use of electronic healthcare record database for safety signal detection in paediatrics. To complement safety surveillance in spontaneous reporting systems and other traditional monitoring systems, initiatives in the United States (US) and in Europe have set up population-based surveillance systems that make use of longitudinal electronic healthcare data. Using a paediatric population that comprised 4.8 million children and adolescents contributing 25.6 million PY of follow-up, we studied what the statistical power of the EU-ADR (Exploring and Understanding Adverse Drug Reactions by Integrative Mining of Clinical Records and Biomedical Knowledge) network is for performing paediatric safety signal detection. The use of medicines in children was rare and was limited to only a few drugs; only 18 represented 50% and 158 drugs covered 90% of the total drug exposure time. The events that were considered to be most relevant for safety monitoring in adults had very low IRs in the paediatric population. The consequence of this combination of rare events and low use of medicines was that the number of medicines with enough exposure to monitor these events was low. For a rare but serious event like anaphylactic shock there were no drugs with enough exposure to study a weak association (RR≥2) and only 20 drugs to study a strong association (RR≥6). For a relatively frequent event such as upper gastrointestinal bleeding (UGIB) there were 5 medicines with enough exposure to study a weak association (RR≥2) and 79 drugs for which an association with a RR≥6, if present, can be investigated. The statistical power of a system like the EU-ADR network is much higher for events more frequently occurring in the paediatric population. In future initiatives to set up drug surveillance systems for the paediatric population using electronic healthcare records, it is very important to choose age-appropriate events and definitions. Inter-continental collaboration will be necessary to gain enough statistical power for paediatric safety surveillance. Signal detection and safety warnings Chapter 10 describes a sequel to the study presented in chapter 9. We compared signal detection for UGIB within a cohort of children and adolescents (aged 0-20 years) from the EU-ADR network with spontaneous reports within VigiBase. Both the number of identified signals and the rate of detected known signals were compared. In total, 142 different signal were identified, of which 17 (12.0%; 41.2% known) in both data sources; 87 in VigiBase only (61.3%; 27.6% known), and 38 in EU-ADR only (26.8%; 13.2% known). VigiBase The signals that were identified only within VigiBase did not have exposure data within EU-ADR or lacked statistical power to detect the event in EU-ADR. Signal detection for children in electronic healthcare record (EHR) data is currently hampered by a lack of statistical power. Enlarging the study population and acquisition of data on in-hospital drug use are essential for improving signal detection within EHR databases. In Chapter 11 we studied the effect of safety warnings on the use of prokinetic (domperidone, metoclopramide and cisapride) medicines in Italy, Spain and the Netherlands. In the last decennium multiple warning have been issued for serious adverse events associated with the use of several prokinetic medicines. Overall a decrease in the utilisation of prokinetic drugs over time was seen in both Italy and the Netherlands while an increase was observed in Spain. Safety warnings had a significant effect on the prescribing in both the Netherlands and Italy. In Spain, the safety warnings did not result in a decreasing prescription rate which shows that mere changes in the Summary of Product Characteristics are probably not enough and higher public notoriety through informative notes and/or the implementation of restrictive measures are needed.

• View at Worldcat

Free Full Text ( Final Version , 5mb )