Introduction

In adults, the risk factors for adverse drug reactions (ADRs) have been defined as old age, polypharmacy and complex underlying disease, but in children, the risk factors (other than polypharmacy) are poorly understood [1, 2]. A factor that has received some attention in this population is the use of off-label and unlicensed medicines [3]. The necessity of using off-label and unlicensed medicines in children is a consequence of how, historically, medicines have been developed and regulated.

The regulation of medicines

Before a new medicine can be approved for use by patients, the manufacturer needs to submit specified information on its quality, safety and efficacy to the relevant national medicines regulatory body. If the new medicine is approved, it is authorised and issued with a Marketing Authorisation (MA). The summary of product characteristics (SmPC) is the important document that is the result of the whole process. It provides vital information for the prescriber that includes the precise indication and dosage of the product, instructions for administration, contraindications, interactions and possible adverse effects. A condition of the approval is that the medicine is only marketed for use under the terms outlined in the MA since these terms reflect the content of the original information submitted by the manufacturer. However, this does not preclude the use of the medicine outside the terms of the MA by individual clinicians.

Off-label and unlicensed medicines

If a medicine has an MA, but is prescribed and/or administered outside the terms of that MA, this is referred to as off-label use or unlicensed use; we use the term off-label in this article. If a medicine does not have approval in the country in which it is prescribed and/or administered, it is referred to as an unlicensed medicine. It is important to note, however, that the exact definitions vary between authors.

Off-label and unlicensed medicine use in paediatrics

Changes to the way medicines are developed mean that every new medicine under development in Europe must have a paediatric investigation plan (PIP) [4], these changes, which were introduced in 2007, are likely to take time to work through into paediatric practice. In the meantime, the use of off-label and unlicensed medicines remains common in this context. A review of studies of off-label and unlicensed medicine prescribing in children showed that it ranged from 3.3 to 56% of prescriptions in community practice, and 36 to 100% in hospital settings [1].

This clearly has implications for the prediction, avoidance, detection, and treatment of ADRs. Safety data for an approved medicine that is being used off-label may not always be relevant or applicable because it relates only to the use of the medicine as specified by the manufacturer. Unlicensed medicines may also not have any safety data detailed in an MA.

Off-label and unlicensed medicine use and ADRs

The results of studies that have examined the association between ADRs and medicines used in this way show much variation. The reasons for this include differences in study design and definitions used. As preparation for a prospective study of ADRs and off-label and unlicensed medicines in children [5], we undertook a comprehensive literature search to evaluate how other investigators have approached the study of off-label and unlicensed medicines. We provide a narrative review of their methodologies and results, and discuss the issues relevant to their interpretation.

Materials and methods

A Medline search of titles and abstracts from 1950 to the present was performed using the search terms unlicensed/off-label/license/licensed/licensing/label/labelled/labelling/approved/approval/unapproved/prescription/prescribed/prescribing/prescribe/prescriber(s)/incorrect AND adverse effects/adverse drug reaction reporting systems/drug therapy/pharmaceutical preparations AND child/child, preschool. An EMBASE search of titles and abstracts from 1980 to the present was also performed using the search terms unlicensed and off-label use AND child AND adverse drug reaction/drug surveillance program. The limits Human and English Language were applied to both searches.

The method used to select papers for inclusion is summarised in Fig. 1, and this was carried out by one reviewer. The titles were screened for reference to off-label and unlicensed medicine use or adverse drug reactions. Papers relating to specific treatments, conditions or reactions were excluded as well as those relating to prescribing and medication errors. Editorials, notes and letters were also excluded. The remaining abstracts (or papers when no abstract was available) were read and excluded if they made no reference to ADRs in the context of off-label or unlicensed medicine use. The studies identified for inclusion in this review are summarised in Tables 1 and 2.

Fig. 1
figure 1

Method of selection for papers for included in this review

Full size image
Table 1 Summary of prospective adverse drug reaction (ADR) studies
Full size table
Table 2 Summary of studies that reviewed spontaneous ADR reports/queries
Full size table

Description of studies

Authors’ definitions

The definitions used by the different authors, together with the setting of the studies varied, and these should be taken into account when interpreting the results of each study.

ADR definition

Only 4 of the 14 studies included gave a definition of an ADR [6–8, 20], each using the WHO definition [9]. Since this definition states that an ADR is “any response to a drug which is noxious and unintended and that occurs at doses used in man for prophylaxis, diagnosis or therapy’” we can assume that the authors who used it excluded errors and accidental poisonings. The results of studies which did not provide a definition of ADR need to be considered in context: some ADRs may have been excluded resulting in an underestimation of ADR prevalence. On the other hand, the definition may also have been expanded to include reactions resulting from prescription or administration errors or from accidental poisonings, which would have resulted in an overestimation.

Off-label medicine use definition

Only 1 of the 14 studies did not provide a specific definition of off-label use [10]. Turner et al. defined six possible types of off-label medicine use: administered at a greater dose than recommended in the MA, at a greater frequency than recommended, for an indication not described, to a child outside the age range specified, via a route not described and when a contra-indication was described [11, 12]. One other study used these exact definitions [13], while six studies used broadly similar definitions [6–8, 14–16]. Neubert et al. [7] used a retrospective approach to classify medicines as off-label or unlicensed. The information available to them did not include an indication for the use of the medicine and the formulation administered was not always recorded. In the absence of an indication, off-label classification was based on age, route and maximum dose for any indication. Gill et al. described only three possible types of off-label medicine use: at different dose, for an indication not described, in a child outside the age range specified, but they did not record their reference source [17]. Jonville-Bera et al. described off-label use as use outside at least one of the recommendations in the SPC, specifically: duration of treatment, dose adaptation, precautions for use, monitoring of treatment, absolute contraindication, indication or route [18]. A Swedish study that used data from the medicines information database included the following in their definition of off-label use: not recommended in children, indication not described, outside the age range specified; they used a national prescribing reference [19]. Schirm et al. defined off-label use as the use of medicines not authorised for use in children or in a child under the minimum recommended age; however, they did not consider dose, frequency, route or indication and did not record their reference source [15]. The most recent study defined off-label use as use in children below the recommended age group listed in the SmPC [20].

Unlicensed medicine use definition

Of the 8 studies that considered unlicensed medicine use in addition to off-label use, 4 provided a definition of an unlicensed medicine [11, 13, 16, 19]. Again, Turner et al. and Clarkson et al. used identical definitions as follows: modification to licensed medicines, e.g. extemporaneous preparations, licensed medicines in a modified formulation manufactured under a special manufacturing license, new medicines available under a special manufacturing license, chemicals used as medicines, medicines used before a license had been granted or imported medicines (i.e. licensed in another country) [11–13]. Kimland et al. categorised any medicine not in their national prescribing reference (Swedish catalogue of medical products—FASS) as an unlicensed medicine [19]. A study carried out in Brazil used the following definitions: medicines contra-indicated for use in children, extemporaneous preparations manufactured or modified by a hospital or nurse or medicines for which safety and efficacy in children was not established [16].

Study methods and results

Prospective studies

A summary of the 7 prospective studies published between 1985 and 2008 included in this review is provided in Table 1. Only 2 of these provided a definition of adverse drug reaction [6, 7], both using the World Health Organisation (WHO) definition [9]. Five studies were set in paediatric hospitals and 2 were in community settings, all employed active patient monitoring and the duration of these studies ranged from 13 weeks to 28 months.

Prospective inpatient studies

Where it was recorded, the ADR incidence in the prospective inpatient studies ranged from 2.53 to 19.9% [11, 17]. One study found a statistically non-significant increase in ADR risk with off-label and unlicensed medicines (RR 1.74, 95% CI 0.89, 3.41, p < 0.106) [11]. Another study found that 11 out of 29 inpatients and 5 out of 12 admissions had ADRs attributed to off-label medicines, but the study size was too small to show a significant association. However, in the same study, a significant association was found between off-label medicine use and ADRs requiring pharmacological intervention (RR 7.0, p < 0.04) [6]. Santos et al. found that off-label medicine use was associated with ADRs (RR 2.44, 95% CI 2.12, 2.89, no p value) [16].

Prospective community-based studies

The ADR incidence recorded by Horen et al. was 1.4% of all patients and was 2% in patients receiving at least one off-label medicine [14]. The other community-based prospective study found an ADR incidence of 11.1% in the study population and an increased relative risk of probable or definite ADRs in patients receiving a total daily dose of medicine above that recommended by the manufacturer (7% vs 4.3%; RR 1.63; CI 1.23 to 2.16; P < 0.001) [10].

Retrospective studies

A summary of the 7 retrospective studies published between 2004 and 2011 and identified for inclusion in this review is provided in Table 2. Two of these studies [8, 20], used the World Health Organisation (WHO) definition of ADRs, while none of the rest of the studies provided a definition. All the studies included ADR reports or queries to either national or regional centres. One study focussed on paediatric reports from hospitals [13]. A limitation that is common to all of these studies is that they are unlikely to provide a representative sample of medicine use and ADR occurrence in a population. The duration of these studies ranged from 5 months to 10 years.

Jonville-Bera et al. reported that 75% of off-label prescriptions resulted in an ADR compared with 59% of the approved prescriptions [21]. Schirm et al. were the only authors who attempted to describe the prevalence of off-label and unlicensed medicine use in the general population—they specify that in their study of Dutch community pharmacies, unlicensed medicines would have only included pharmacy preparations. They found that 24% of medicines implicated in spontaneously reported ADRs were off-label and 1.9% were unlicensed [15]. Four studies reported that between 17 and 42.4% of spontaneous ADR reports were associated with off-label medicine use [8, 13, 18, 20]. The study that considered medicine information queries relating to ADRs found that 27% involved off-label medicines and 17% involved unlicensed medicines [19]. A study of spontaneous paediatric reports from hospitals observed that 6 out of 10 medicines implicated in fatalities were off-label or unlicensed [13].

Concluding remarks

Although the results of previous studies have indicated that there may be some association between off-label and unlicensed medicine use and ADR risk [11, 16], there is still a lack of clarity. This stems from the fact that some of these previous studies were small, they employed different methodologies as well as inexact and varying definitions.

There is a need for further research in this area, in particular to ensure that studies are of adequate power to answer some key questions. One of these questions is whether the risk factors for off-label prescribing and those for ADRs converge. The risk factors for ADRs have been discussed above. Off-label and unlicensed medicine use has been found to be more prevalent in certain specialities, e.g. ophthalmology, cardiology and dermatology, in infants, in children who consult their general practitioner more often, receive more prescriptions or have more specialist referrals than other children [22, 23].

Recommendations

Further studies in this area are required and they need to be designed with attention to detail to ensure that the results are reliable, interpretable and useful. Clear definitions of ADR, off-label and unlicensed medicine use should be provided. In our view, the definitions originally provided by Turner et al. [12] are the most comprehensive and unambiguous definitions published. This is borne out by the fact that they have been used by a number of other authors [6–8, 13–16]. That is not to say that these definitions may not need to be modified to suit the study setting; indeed some of the authors who employed them made modifications. If off-label and unlicensed medicine use is categorised using these definitions it should be possible to use the results of future studies to test hypotheses on the reasons why off-label and unlicensed medicine use may be implicated in ADRs, for example: are certain types of off-label or unlicensed medicine use more likely to be implicated in an ADR than others?