Objective To elucidate the current situation of randomised drug trials involving the paediatric population.
Methods Systematic review of paediatric randomised controlled trials of medicinal products published in 2007. Three major databases were searched with validated search strategies; Medline, Embase and Cochrane Central Register of Controlled Clinical Trials. Data was collected on the location, participants, class of drug and methodological quality of the trials.
Results Six hundred and four trials were found involving more than 100 000 paediatric participants. Only about a quarter (146, 24%) were conducted in low and lower-middle income countries. Few studies (42, 7%) were performed in neonates. Many trials recruiting both adult and paediatric patients inadequately describe the characteristics of the paediatric participants. The most studied areas were nervous system (155, 26%), anti-infective (101, 17%) respiratory (74, 12%) or antiparasitic (45, 8%) drugs. A high proportion of the studies (36%) used an inactive placebo as the comparator. Paediatric randomised drug trials performed in low and low-middle income countries were of lower methodological quality (mean Jadad score 2.90 vs 3.27, p<0.01), studied more antiparasitic and anti-infectives (47% vs 16%, p<0.01) but fewer reported that ethical approval was obtained (83% vs 93%, p<0.01), compared to those conducted in high or upper-middle income countries.
Conclusions There are a significant number of randomised controlled drug trials involving children taking place throughout the world. To develop the evidence base for safe and effective medicines for the benefit of the whole paediatric population, high quality and ethical clinical trials should involve a wide range of children.
Statistics from Altmetric.com
In the past decade, the medical community has seen major developments in the effort to provide children with medicines that have undergone rigorous scientific evaluation of their efficacy and safety.1 Legislation in the USA has been seen to increase the number of clinical trials in paediatric populations.2 There is now European legislation with similar financial provisions to encourage clinical studies of medicines in children.3
There is growing realisation that the problem is a global one.4 The WHO launched a global campaign to ‘make medicines child size’ in December 2007, spearheading efforts to increase the development and access to childspecific medicines.5 A major thrust of the campaign is to tackle the priority research gaps, where medicines do not exist, or safety and efficacy are not known in children. The priority medicine areas identified were the major infectious diseases; HIV/AIDS, malaria, pneumonia, tuberculosis and diarrhoea. The majority of children with these diseases live in low and lower-middle income countries (LMIC). However, previous reports have documented that few paediatric clinical trials involved children in these countries.2 6
What is already known on this topic
There have been major developments to encourage clinical drug studies in children including legislation in the USA and now Europe.
WHO launched the ‘make medicines child size’ campaign in December 2007 to improve access and availability to safe, effective child-specific medicines worldwide.
What this study adds
There are a significant number of randomised drug trials involving children taking place worldwide, and many appear to address the appropriate clinical areas.
Relatively few randomised controlled drug trials involve children in LMIC, and neonates.
In this context, a review of randomised drug trials involving children published in 2007 was performed to provide baseline data on the children involved, the medicines studied and how many and where the trials were conducted globally.
A descriptive study involving a literature review of randomised clinical trials of medicines involving children published in 2007 was conducted. Three major databases namely Medline, Embase and the Cochrane Central Register of Controlled Clinical Trials (Central) were searched.
Medline was searched electronically via the Ovid gateway using the CRD/Cochrane Highly Sensitive Search Strategy (2005 revision),7 with validated age limits.8 Results were limited further to therapeutic drug trials only.
Embase was also searched electronically via Ovid using a validated search strategy developed by the Hedges Team.9 As there has been no validated age-specific search strategy developed for Embase, Ovid's own age group classification was used to limit the results to paediatric age groups.
Central was searched via the Cochrane library website. Age-specific keywords were derived from the BestBETs group search strategy to identify clinical trials involving paediatric age groups.10 All searches were limited to articles published in 2007.
The resulting abstracts were compiled in individual libraries using the software Endnote (V.X.0.X1). Intrasearch and intersearch duplicates were identified and removed.
The following inclusion criteria were applied to the abstracts to identify relevant studies: randomised controlled trials only, randomised participants included any of the International Conference for Harmonisation paediatric age groups,11 the main intervention studied is a medicinal product.
Each and every abstract of the retrieved reports were read and evaluated. When there was uncertainty, the full paper was obtained to determine inclusion. Full papers were sourced for all abstracts deemed appropriate for inclusion. The full papers obtained were then read and data was entered into a database using SPSS V.15. There was no restriction for language.
Data was collected on the setting of the trials. The recruitment location of each trial was determined and categorised according to the World Bank income level (2007 Gross National Income per capita)12 and the United Nations Development Programme Human Development Index (2007 human development index (HDI)).13 Cross-border or international trials were identified. For each trial, institutional setting and number of centres were noted.
The number of children taking part and their age groups were classified. The main drug class studied for each trial was coded according to the WHO Anatomical Therapeutic Chemical classification system.14 Information regarding route of administration and type of comparison was collected.
The methodological quality of each randomised trial was evaluated according to the Jadad scoring system, a validated five-point system widely used in evidence based reviews.15 Each study was given a Jadad score. Whether each trial documented ethical approval, informed consent and child assent was also determined from the reports. Financial support declarations were also identified and categorised.
For statistical analysis, means were compared using the unpaired t test while differences in proportions were compared using Fisher's exact test with two-tailed p values.
The search strategies yielded 15 577 abstracts from the three databases. After duplicates were removed and inclusion criteria applied to the abstracts, 582 full reports of trials were obtained and entered into the database. Twenty-two non-English reports fitting the inclusion criteria were also identified. Data was extracted from the English language abstracts of these reports. Please see figure 1 for flow of trial reports.
Approximately one in four trials (146, 24%) was conducted in LMICs (table 1). The majority of the studies (352, 58%) originated from high income countries. A similar trend was seen when the settings were categorised according to HDI status. In all, 153 (25%) and 28 (5%) studies were performed in medium and low HDI countries respectively. Of the trials, 392 (65%) were conducted in high HDI countries. Thirty-one (5%) trials recruited patients across HDI groupings.
Taking into account the wide variation of healthcare systems between countries, a simple classification system was used for setting. There were 307 (51%) studies performed in outpatient departments or in general practice. In 219 (36%) trials the patients were hospital inpatients. Another 69 (11%) studies were done in the community for example in schools, villages or towns.
In all, 323 (54%) of the trials were single-centre studies while 268 (46%) were multi-centre studies. There were 96 (16%) cross-border studies that recruited in more than one country.
More than 100 000 paediatric patients took part in randomised controlled drug trials published in 2007. The number of paediatric patients in each study ranged from two to 8352 (median=90, interquartile range=41 to 210 patients). In all, 428 (71%) studies recruited paediatric age groups only, while 176 (29%) studies recruited both adult and paediatric patients. In 143 (24%) of the studies, the number of children participating was not able to be ascertained because these mixed-age trials did not report the specific number of participants by age group (table 2).
Forty-two (7%) of the trials recruited neonates. Twenty-six involved preterm neonates and 23 full term neonates (seven studies involved both). The majority of the studies included children (2–11 years) or adolescents (12–16 years) (figure 2). Approximately one in four studies included infants (28 days to 23 months).
Drugs and diseases
Table 3 shows the main drug therapy studied in relation to the WHO Anatomical Therapeutic Chemical drug categories. The most common categories were nervous system followed by systemic anti-infectives and respiratory drugs. These three areas involved over 50% of all the trials. The greatest number of nervous system drugs studied were anaesthetic agents (56 trials) followed by analgesic drugs (33 trials) and stimulant drugs used in attention deficit hyperactivity disorder (29 trials). Vaccines (51 trials) and antibacterial drugs (32 trials) formed the majority of the systemic anti-infectives. Most of the respiratory drugs studied were drugs for obstructive airway diseases (54 trials) consisting mainly of anti-asthmatic medications.
Trials of antiparasitic medications and systemic anti-infectives (69/146 trials vs 68/418, p=0.001) were more frequent in LMIC, compared to upper-middle and high income countries. Frequencies of nervous system drug trials were similar (33/146 vs 118/418, p=1.0) while significantly more respiratory drug trials were conducted in upper-middle and high income countries (56/418 vs 3/146 trials, p=0.0004).
The major routes of administration were oral (39%), intravenous (20%), intramuscular (8%), local pulmonary and topical dermal (7% each). There were 242 (40%) active-comparator trials where the main drug therapy studied was assessed against another drug. A similar number of trials, 213 (36%), used a placebo control. Seventy-nine (13%) of the studies were trials comparing different dosing regimens or formulations of the same drug. The remainder consisted of trials with untreated controls, comparing different routes of administration or trials comparing drugs to non-pharmacological therapies. The vast majority of the trials (91%) utilised a parallel-group design while the remainder were crossover studies.
The mean Jadad score for all the studies was 3.22 (SD=1.31). Based on the study design, all of the studies included in the database were described as randomised. Therefore all the studies scored at least 1 point on the Jadad scoring system. For the other items, the scoring rate ranged between 49% and 67%. Table 4 shows the scoring for the each of the items.
Trials conducted in high and upper-middle income countries scored significantly higher Jadad scores compared to those conducted in LMIC (mean=3.27 vs 2.90, p=0.003).
Of the reports, 91% described that the study received approval from an ethics committee or an institutional review board. In all, 92% of the studies declared that informed consent was obtained. Assent from the child or where consent was obtained directly from the paediatric patient was described in 128 (22%) of the trials. In studies that included adolescents 12–16 years old, only 109/366 (30%) documented that assent was sought. A lower proportion of trials from LMIC mentioned that ethical approval (82.6% vs 92.5%, p=0.0015) for the study was obtained.
Of the total number of randomised drug trials in the database, 39% lacked an adequate and transparent declaration of funding source for the study. In the reports that did acknowledge funding source, pharmaceutical companies constituted the majority of trial sponsors (30%). Fourteen per cent of the trials were mainly funded by academic funds, 11% by health or governmental authorities and the remaining studies attributed their funding to charities or foundations.
More than half of studies (75/146, 51%) conducted in LMIC did not declare funding source. Of the 71 studies that did, academic (31%) and charitable funding (25%) were the main sources mentioned. Over half (40/71, 56%) declared an overseas organisation as the main trial sponsor.
It is reassuring that large numbers of randomised controlled drug trials are taking place globally. This should increase the evidence base for safe and effective use of paediatric medicines around the world. The disease burden in children lies mostly in LMIC, where it is estimated that 99% of all child deaths occur.16 17 Although fewer paediatric randomised drug trials were conducted in LMIC, the disease areas in which these trials took place seemed appropriate considering the majority of the trials studied anti-infective and antiparasitic drugs. Trials conducted in high and upper-middle income countries also seemed to target the diseases appropriate to their setting, particularly asthma medications.
The development of safe and effective child-specific treatments requires high quality clinical trials.18 Furthermore conducting clinical trials in poor countries may bring ancillary benefits to the population.19 There is mounting evidence of the globalisation of clinical research where clinical trials are increasingly being performed away from developed countries.20 However, this trend is mainly due to economic factors motivating the pharmaceutical industry.21 We found that a significant number of trials in LMIC reported that their funding came from abroad. Therefore paediatric health professionals, researchers and regulatory bodies should play an active role in ensuring clinical trials performed in LMIC are ethical and of high quality.
Ethical concerns regarding clinical trials in developing countries or LMIC has been frequently discussed.22 23 A major issue is ethical oversight of research involving human subjects. A previous study found that only 56% of researchers in developing countries in Asia, Africa and South America reported that their research had undergone ethical review.24
Another study found that only 10% of randomised controlled trials performed in China reported that ethical approval had been obtained, and only 18% reported on informed consent.25 The majority of paediatric clinical trials in this review reported both ethical approval and informed consent. However, more attention should be paid to paediatric drug trials in LMIC so that ethical approval and informed consent is universally documented. Another noteworthy finding was the low proportion of studies documenting that assent was obtained, especially from adolescent participants where only 30% of the studies documented this process. There is growing opinion that assent should be sought from paediatric participants in clinical trials, especially older children and adolescents.26 27
Previous concerns have been raised that studies conducted in children have been in drugs that confer high financial returns rather than drugs that benefit children the most.28 It was noted that from the paediatric studies submitted to the US Food and Drug Administration between 1998 and 2004, the highest numbers were in psychotropic drugs (31), antihypertensives (22) and studies of conjunctivis/rhinitis drugs (18). Only 11 studies of HIV drugs, nine of drugs for respiratory infection and six studies on antimalarials were submitted.29 The large numbers of nervous system drugs in this review was interesting, particularly since they were mostly anaesthetics, analgesics and attention deficit hyperactivity disorder drugs. Anti-epileptics only account for slightly over 10% of the nervous system drugs studied despite reported figures that state 60–90% of people with epilepsy worldwide are untreated or inadequately treated, including vast numbers of children.30 31
The many trials of anti-infective, antiparasitic and respiratory drugs were encouraging as they broadly correspond with the major target areas highlighted by the WHO ‘make medicines child size’ campaign. The results also show that trials of antiparasitic and anti-infective drugs constituted the majority of studies conducted in LMIC. However there were only seven trials that studied antivirals and antimycobacterials, which was surprising considering that HIV/AIDS and TB cause almost half a million child deaths annually.17
The high proportion of placebo controlled trials was interesting considering the increasing concern that these trials may be unethical whenever a comparator drug exists and therefore potentially effective treatment will be withheld.26 Further work is needed to examine the appropriateness of the placebo controlled trials identified in this review. We found few studies of medicines in neonates. Drugs used in neonates are more likely to be unlicensed or off-label among the paediatric age groups.32 Furthermore, neonates appear to suffer high rates of adverse drug reactions.33,–,35 More research is needed to evaluate the appropriate amount and clinical areas with regard to clinical trials involving neonates.
Another concern was that trials recruiting both children and adults often inadequately describe the characteristics of paediatric participants. This is essential to report because of the potential differences in pharmacokinetics and pharmacodynamics between paediatric and adult participants.36
The study suffered from some limitations. Although the search strategies did not limit for language, the databases searched were heavily biased to English scientific literature. An analysis of the 498 English abstracts of non-English papers only found 22 paediatric randomised drug trials.
Relevant studies, especially those reported fully in languages other than English, may be published in local journals not indexed in the databases searched. Another weakness relates to publication bias. It is acknowledged that not all paediatric randomised drug trials conducted would be published in peer-reviewed journals. However, it is felt that this wide-ranging review using sensitive and validated search strategies would reflect the overall situation of paediatric randomised drug trials.
In conclusion it is encouraging that a large number of trials have been identified, taking place all over the world with over 100 000 children involved. However, fewer trials took place in LMIC or studied neonates. High quality, ethical clinical trials involving a wide range of children are essential to improve child health on a larger scale. The ‘make medicines child size’ campaign by WHO should receive strong and continuing support from the medical community to benefit children worldwide.37
KNBNA holds a scholarship granted by the Ministry of Higher Education, Malaysian government.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.