Antimicrobials are among the most commonly prescribed drugs in paediatrics. However, up to 50% of in-hospital antimicrobial prescriptions are inappropriate.1 The relationship between inadequate use of antimicrobials and the emergence of bacterial resistance has been clearly established, resulting in increased mortality and costs.2 Antimicrobial stewardship programmes (ASP) were successfully implemented to optimize antimicrobial use in hospitalized patients.3 Though, development of ASP in paediatrics has been limited, due in part to the lack of a standardized method for comparing antimicrobial use.4

Defined Daily Dose (DDD) is one of the established metrics used by ASP, allowing the assessment of antimicrobial consumption. The World Health Organization (WHO) expresses DDD as the average standard daily dose of a drug used in a 70kg adult for the most common indication.5,6 However, the validity of DDD WHO definition is questionable in hospitalized children, in which dosing is based on body weight.1,5 This study aims to establish a methodology for antimicrobial DDD measurement in the paediatric population.

A total of 40,575 patients were included: 23,877 (58.8%) males, 16,698 (41.2%) females. The median age was 4.17 (IQR: 1.36–8.98) years old for boys and 4.81 (IQR: 1.42–9.60) for girls. The 50th percentile of weight for that age was 17.08kg (males 16.6kg and females 17.8kg). The selected antimicrobials with their respective calculated DDDs and agreement percentages are shown in Table 1 (intravenous route) and Table 2 (oral route). In 9 of 52 antimicrobials, a second round was necessary for dose agreement, achieving it in 4 of them. For oral amoxicillin, amoxicillin-clavulanic acid, ampicillin, cefuroxime and intravenous ceftriaxone; dose was agreed using Pediamecum data. Total agreement percentage was 77.9% (360/462): 79.3% (242/305) and 75.2% (118/157) for intravenous and oral antimicrobials, respectively.

Paediatric DDDs for antimicrobials have been designed in this multicentre study. The best metric for aggregate antimicrobial consumption evaluation in paediatrics has not yet being defined.4,10 DDD measurement in children is limited due to weight variability, thus, other methods like Point Prevalence Surveys (PPS) have been proposed.2,11,12 PPS can assess antimicrobial consumption in short periods of time basis, using retrospective, prospective or mixed designs.10,12 The Antibiotic Resistance and Prescribing in European Children Project (ARPEC) and the European Surveillance of Antimicrobial Consumption (ESAC), have used this metric to monitor antimicrobial consumption in children.10,11 However, PPS specific time-point data is susceptible to case-mix complexity, seasonality and variability of the sample.

Current recommendation for paediatric ASP metric is Days of Therapy (DOT).5 This metric counts the number of days a patient receives antibiotics, avoiding weight variability. Nonetheless, DOT measurement poses problems in the setting of drugs with a long half-life, variable pharmacodynamics, or in patients with renal failure.5 In addition, DOT favours patients who receive less antimicrobials even if that involves the use of broader spectrum antimicrobials,5 while the optimal length of treatment for most of paediatric infections varies widely.12

Few studies have used the DDD methodology in children.4 A recent systematic review found up to 26 distinct measures across 79 studies including paediatric patients (DDD or similar metrics in 38 studies).4 Limitations were found for all metrics and authors conclude that the best metric remains to be identified.4 In our study, weight variability was overcome by selecting the average weight for the obtained sample. The availability of a normalized DDD value could allow its application for ASP monitoring in paediatrics. Nevertheless, the average weight obtained in our sample, does not allow comparisons with other series that may report a different median age and consequently a different average weight. Future strategies could include constructing confidence intervals for the calculated DDDs or designing specific DDDs for paediatric age subgroups of interest (neonates, infants, children, adolescents).

Centre participation was high for the selected dose agreement method, therefore, the lack of dose values for some antimicrobials does not reflect data loss, it could be due to the variability of antimicrobial use in the included centres. Thus, it is possible that some agreed DDDs of occasionally used drugs are less reliable, as well as for those antimicrobials for which there were more discrepancies between centres. Nevertheless, dose agreement data was high for most of the antimicrobials included, and further confirmation would be expected when this tool undergoes validation.

Of note, 4 out of 52 antimicrobials had a greater paediatric DDD when compared to adult DDD. For amoxicillin and liposomal amphotericin B, the DDD value established by WHO is smaller than the dose used in clinical practice according to the specific antimicrobial drug label instructions.13 For cefepime and oral fosfomycin the agreed doses where greater in paediatrics probably because these antimicrobials are used mostly for infections that require higher doses in children.14

Main limitations of our study include the lack of validation of this strategy and low external validity of the method. To overcome this, a new multicentre study has been projected to validate this tool in real-life patients who receives antimicrobial prescriptions under usual clinical practice. Currently, the project is looking for internal validity in our country including representative centres of several Spanish regions. If good results are obtained, further international multi-centre studies will be proposed to standardize this metric.5,6 Although standardized Spanish weigh for age and sex charts exist,15 WHO charts were used with the aim to provide a tool potentially applicable in different populations.

To the best of our knowledge, this study provides one of the largest sample size to define specific DDD for paediatric population, in a multicentre setting. Neonatal patients were excluded, as they have different pharmacokinetics and pharmacodynamics compared to paediatric patients. A parallel project is underway to design neonatal DDD applying a similar methodology.

In conclusion, this study has designed a suitable paediatric DDD for antimicrobial consumption assessment. Further validation of this tool standardizing the age ranges is required in future studies, as this metric could help to design paediatric ASP strategies to prevent the emergence and dissemination of antimicrobial resistance in children.

No specific funding was received for this article.

The authors declare no conflicts of interest.