Systematic review of paediatric studies of adverse drug reactions from pharmacovigilance databases.

OBJECTIVE
To perform a systematic review of studies describing paediatric adverse drug reactions (ADRs) conducted from national pharmacovigilance databases.


METHODS
A systematic literature search of studies describing results for paediatric ADRs from national pharmacovigilance databases was performed. PubMed database, Embase and MEDLINE were searched up to March 2015. The descriptive studies included were analysed for country of origin, reporters, and ADR reporting rate, drugs, ADRs and number of fatalities.


RESULTS
20 studies were identified. Doctors were the largest group of reporters in all the studies, and with more consumer reports seen in USA. The studies ranged from 3 - 37 years. The highest ADR reporting rate was 1458 reports per year per million children in Cuba. Antibiotics and vaccines were the most frequently reported drugs, in almost all the studies. The most frequent ADRs were skin and nervous system disorders. The highest proportion of fatalities and serious reports was from North America. Drugs used for treating attention deficit hyperactivity disorders (ADHD) and isotretinoin were the most frequently reported drugs for ADRs in North America.


CONCLUSIONS
There were geographical differences in drugs responsible for ADRs and their seriousness, especially in North America. Very few studies were conducted in Asia and Latin America, none were found from Africa.


Introduction
Adverse drug reactions (ADRs) are a major problem for children in all countries [1]. One in 10 children in hospital in a high-income country (HIC) will experience an ADR [2]. A study in Cuba suggested that at least 1 in 500 children will experience an ADR each year [3]. National pharmacovigilance databases have been useful in detecting signals. They have also been helpful in describing the type of ADRs experienced by children in a country and the drugs responsible [4]. The majority of the published studies describing ADR data for children have originated from HICs in Europe. The most frequently described ADRs in children in Europe are rashes, headache, pyrexia, and gastrointestinal disorders. The medicines responsible were antibiotics and vaccines.
Antibiotics and vaccines are the most widely used medicines in children worldwide. However, there are significant differences in the use of antibiotics in different countries. Additionally, there are major differences in the use of medicines in children, even in Europe [5,6]. Therefore, one may see differences in the types of ADRs experienced. Additionally, the prevalence of diseases will be different in low-income and lower-middle-income countries. One would therefore expect the ADRs to be different. Antimalarials, for example, are infrequently used in children in Europe. Many ADRs are preventable due to inappropriate prescribing. It is only by identifying the drugs responsible for ADRs that one can determine areas where prescribing may be inappropriate. We therefore decided to look at all publications reporting ADRs in children from national pharmacovigilance databases to explore the differences between ADRs and drugs responsible in different countries.

Methods
A systematic literature search of published studies describing ADRs from national spontaneous reporting pharmacovigilance databases was performed. In order to capture the relevant articles, the following terms and/or keywords assigned to articles which studied ADRs from databases were used: adverse drug reaction*, adverse effect*, adverse event*, side effect*, pharmacovigilance, spontaneous reporting, ADR reporting system, ADR reporting, drug surveillance program, and adverse event reporting. The databases used were PubMed, Embase, and MEDLINE. The PubMed database was searched from when the World Health Organization (WHO) Collaborating Centre for International Drug Monitoring in Uppsala was established in 1978 to March 2015. The WHO center receives individual case safety reports of suspected ADRs from national pharmacovigilance centers around the world [7]. Embase and MEDLINE were searched from 1980 to March 2015 and 1947 to March 2015, respectively. In addition to these searches, the reference lists of publications were also searched. The search strategy is as shown in Table 1.
Articles which described ADRs from national, international, and regional pharmacovigilance databases were considered.
Studies describing specific ADRs and drugs, vaccines, specific treatments, and procedures for disease conditions in pharmacovigilance databases were excluded. All ADR studies outside of national, international, and regional pharmacovigilance databases were also excluded.
All the included articles were considered by country of study, population, duration of study, number of patients/ reports, notifiers, types of drugs and ADRs, death reports, and whether the data used for the studies were from a regional, national, or international ADR database. The articles identified were separated into those reporting on the general population and pediatric population. Only the pediatric studies were analyzed.
The ADR reporting rate for studies where this was not reported was calculated as follows: Reporting rate = annual average number of reports/population × 1 million The population estimates were taken from United Nations Population Division, World Population Prospects: The 2012 Revision [8].

Results
The literature search produced a total of 6720 articles from PubMed, Embase, and MEDLINE. When the inclusion and exclusion criteria were applied, 41 articles were identified, made up of 20 for pediatric studies and 21 for general population studies ( Figure 1). Only the 20 pediatric studies were analyzed.

Summary of findings
The 20 studies reviewed included 3 international, 12 national, and 5 regional studies. The majority of the studies were conducted in Europe. Four were from North America, two from Latin America, one from Asia, and none from Africa.
Eight of the studies excluded vaccines (including all four from North America) ( Table 2). The majority of the studies grouped ADRs in system organ class (SOC) and suspect drugs in anatomical therapeutic class (ATC) or drug class, and presented them in percentages of first, second, and third most frequently reported. The type of notifiers and death reports were also presented in percentages.
The period of time data was collected for the studies ranged from 3 to 37 years. The 37-year data described fatalsuspected ADRs in UK between 1964 and 2000. For the studies which described seriousness of the events, the percentages of ADRs or reports ranged between 2% and 68% of the total number of ADRs or reports. The highest proportion of serious reports was from North America (42-62%). Four studies described reporting rates. The reporting rates for the other studies were calculated using an estimate of the childhood population for the year of study. The highest ADR reporting rates of 634 and 1458 reports per year per million children were described in regional studies conducted in Cuba in 2010 and 2012, respectively.
Nine studies described notifiers of ADRs (Table 3). Doctors were the largest group of reporters in eight of the studies. Pharmacists and other healthcare professionals (HCPs) (including nurses) also submitted reports in most countries. The largest proportion of reports from consumers was from the USA.

Reported ADRs
The majority of the studies reported skin disorders (rash and urticaria) as the most frequent ADRs (Table 4). Other common ADRs were nervous system disorders (headache, dizziness, and drowsiness) and pyrexia/fever. The less commonly reported ADRs were convulsions, diarrhea, and agitation. One study [19] described only fatal reports, and the most frequent fatal ADR in this study was hepatic failure. The majority of the studies which ranked the frequency of ADRs described them in SOC. Only a few studies described ADRs in lowest level terms.

Reported drugs
The most frequently reported drugs are shown in Table 5. Antibiotics and vaccines were the most frequently reported drugs in almost all the studies identified in Europe, Latin America, and Asia. Amoxicillin was the most frequently reported individual antibiotic (where this was stated) apart from in one Italian study where it was second after amoxicillin/clavulanic acid and in one Chinese study where cefuroxime was the most frequent.
In contrast, in North America, drugs used for treating attention deficit hyperactivity disorders (ADHD) (methylphenidate was named in one study) and isotretinoin were most frequently reported. There was no uniformity in the way frequency of drugs reported were described in the studies. In most studies, ATC was used and in others, specific drugs were described. Drug surveillance program 8 Adverse drug reaction reporting 9 Adverse event reporting 10 6 OR 7 OR 8 OR 9 11 5 AND 10

Deaths reported
The studies conducted in North America reported the highest fatality rates, which ranged from 3.4% to 13.0% (Table 6). These rates contrasted sharply with the rates reported in Europe, which ranged between 0.1% and 2.2% with a mean rate of 0.8%. In the UK study [19] which described 37-year data, 0.8% fatality was recorded.

Discussion
Systemic anti-infectives and vaccines were the group of drugs most frequently associated with ADRs in studies in Europe, Asia, and Latin America. This reflects the widespread use of anti-infectives and vaccines in children [27,28]. Many antiinfectives, however, are prescribed inappropriately [5,6]. Reducing inappropriate prescribing would reduce the number of ADRs. In contrast, in North America [21][22][23], drugs used for treating ADHD and isotretinoin were most often associated with the ADRs. These findings suggest that the epidemiology of the use of medicines in children in North America may be different to other countries. The diagnosis of ADHD in children visiting outpatient clinics in the USA has been on the rise since 2000 [29]. Interestingly, antiepileptic drugs were the third most frequently reported group of drugs in most countries, including North America. The most frequently reported ADRs were skin disorders and pyrexia. The proportions of deaths in the reports were higher in North America compared with those in Europe, Asia, and Latin America. This may reflect differences in the use of medicines and also differences in attitudes toward reporting. This is an area open to further research.
The countries which have established programs of ADR reporting in their health-care system are likely to have more reports in national ADR databases. The size of the country, population, and duration of time of data collected for the study may also influence the number of reports in the database. The combination of these factors may have contributed to the higher number of reports or rates of reporting per population observed in the national studies conducted in UK [14,19], Denmark [15], Sweden [16], and USA [21,24]. The high reporting rates recorded from the Paediatric studies 20   studies in a regional ADR database in Cuba [3,25] was attributed to a special educational program on pharmacovigilance for HCPs. The high number of reports received per year in these countries is an indication of success of the reporting scheme. Even though most of the studies in Europe and North America showed evidence of efficient reporting scheme, not all the HPCs were actively engaged in reporting ADRs. The results showed physicians and doctors submitted more reports than pharmacists and nurses. Other HPCs, especially nurses, could be encouraged by giving them focused education to detect and report ADRs [30], due to their unique role in observing patients' signs and symptoms. This approach has been successful in Sweden [31] and Latin America [25].
This study has several limitations. Very few studies were conducted in Asia and Latin America; therefore, comparison of reported ADRs and drugs between the continents was not possible. No study was identified from Africa. Since the search was performed, however, we have published two studies describing ADR reports for children in Nigeria and Ghana [32,33]. There was no uniformity in the reports in terms of ADRs and drugs description. Authors described ADRs and drugs differently in the studies. Some of the studies described ADRs in SOC and others in lowest level reaction terms. The drugs were also described differently in ATC and specific drugs/active substances. All these made comparison of the studies difficult.

Conclusions
The majority of the studies reviewed were from Europe and North America, no study was identified from Africa. ADR reporting rates are higher in Europe, North America, and Latin America compared to Asia. Physicians and doctors reported more ADRs compared to pharmacists and nurses.
Vaccines and anti-infectives were most frequently associated with ADRs in children, with the exception of North America. The reported fatality rate was higher in North America.

Expert opinion
ADRs are underreported worldwide. The highest reporting rate was 1458 reports per year per million children in Cuba. Reporting rates in Europe (the continent with most studies) were considerably lower. Education about ADRs and their reporting is needed in Europe. Antibiotics and vaccines were the most frequently reported groups of drugs in most studies. They are also the most frequently prescribed groups of drugs to young children. Antibiotics are often prescribed inappropriately and efforts to improve rational prescribing are needed. In contrast, in North America, drugs for ADHD were the most frequently reported drugs. This may be related to the widespread use of these drugs in North America. More reports of  pharmacovigilance are needed from low-and middleincome countries in Africa, Asia, and Latin America.

Funding
This paper was not funded.