Hospitalised neonates in Estonia commonly receive potentially harmful excipients

Background Information on the neonatal exposure to excipients is limited. Our aim was to describe the extent of excipient intake by Estonian neonates; to classify the excipients according to potential neonatal toxicity and thereby to measure the extent of exposure of neonates to potentially harmful excipients. Methods A prospective cohort study that recorded all medicines prescribed to patients aged below 28 days admitted to Tartu University Hospital from 01.02-01.08 2008 and to Tallinn Children’s Hospital from 01.02- 01.08 2009 was conducted. Excipients were identified from Summaries of Product Characteristics and classified according to toxicity following a literature review. Results 1961 prescriptions comprising 107 medicines were written for 348/490 neonates admitted. A total of 123 excipients were found in 1620 (83%) prescriptions and 93 (87%) medicines. 47 (38%) of these excipients were classified as potentially or known to be harmful to neonates. Most neonates (97%) received at least one medicine (median number 2) with potentially or known to be harmful excipient. Parabens were the most commonly used known to be harmful excipients and sodium metabisulphite the most commonly used potentially harmful excipient, received by 343 (99%) and 297 (85%) of treated neonates, respectively. Conclusions Hospitalised neonates in Estonia are commonly receiving a wide range of excipients with their medication. Quantitative information about excipients should be made available to pharmacists and neonatologists helping them to take into account excipient issues when selecting medicines and to monitor for adverse effects if administration of medicines containing excipients is unavoidable.


Background
Excipients are essential components of medicinal products required for manufacturing processes to assure several properties such as solubility, bioavailability and stability of the final dosage form. Concerns about the safety of pharmaceutical excipients are growing due to the increasing number of adverse reports, especially in neonates [1][2][3][4]. According to regulatory requirements, excipients have to be appropriately evaluated for safety [5][6][7]. Similar to active pharmaceutical ingredients, in most instances the safety data of excipients is based on adult exposure. Thus, information about their acceptability and safety in relation to the age and development status of the child is often lacking [8].
Neonates are the most vulnerable patient population when adverse reactions of excipients are considered. This is mainly due to organ immaturity and differences in pharmacokinetic (PK) and pharmacodynamic (PD) profiles compared to adults [9]. The inclusion of inadequately studied excipients in products used in neonates has resulted in tragedies such as the E-ferol incident where over 30 infants died after receiving an intravenous vitamin E product containing polysorbate 80 [10]. The potential for dose-related adverse reactions of excipients are of particular concern in preterm infants due to the immaturity of hepatic and renal functions [11,12].
The use of potentially toxic excipients in medicines given to neonates is not rare [13] as they are present in many commonly used drug products [14,15]. However, neonatal exposure to excipients is still poorly studied. Previous studies have been selective in terms of populations, indications or excipients. For example, the studies have included premature neonates only [13] or have been restricted to medicinal products used in gastroenterology [14] or to excipients known to be toxic (sodium benzoate, propylene glycol, methyl-and propyl parahydroxybenzoate, saccharin sodium, benzyl alcohol, benzalkonium chloride, polysorbate 80 and ethanol) [13,16]. We are not aware of studies looking at frequency of excipient use in an unselected cohort of hospitalised neonates.
Our objectives were first, to record the frequency of using potentially harmful excipients in hospitalised neonates; second, to classify the excipients used in neonatal wards in Estonia into categories according to the possible toxicity to neonates; and third, to describe how many neonatal prescriptions included potentially harmful excipients.

Methods
This study used a cohort design recruiting in two centres over two non-overlapping but equivalent periods. A total of four wards in both paediatric hospitals of Estonia participated. All medicines prescribed to hospitalised neonates with postnatal age (PNA) below 28 days were recorded in Tartu University Hospital (TUH) from 1 st of February to 1 st of August 2008 and in Tallinn Children's Hospital (TCH) from 1 st of February to 1 st of August 2009 [17]. The following information was extracted from the hospital records on twice weekly visits: demographic data (gestational age [GA], birth weight, gender, date of birth, PNA), admission and discharge dates and prescriptions for all medicinal products (International Nonproprietary and product names, formulations). The dose and frequency of medicines were not recorded for the purpose of this study and the use of standard intravenous replacement solutions, nutritional and technical products (including contrast agents), basic creams and ointments, parenteral nutrition solutions, vaccines and vitamins were excluded, although we acknowledge that some of these products may also contribute to the general excipient exposure.
The data from the neonatal wards at both hospitals were pooled and patients categorised based on the GA as preterm (<37 weeks) and full term (≥37 weeks) neonates.
The excipients present in the prescribed medicines were identified from their Estonian Summaries of Product Characteristics (SPC) and if the drug product was not registered in Estonia in September 2009, from the package insert leaflets. The excipients were divided into four categories as detailed in Table 1. The following literature sources were used for classification: Rowe's Handbook of Pharmaceutical excipients 6 th ed. [18], European Commission guidelines on the excipients in the label and package leaflet of medicinal products for human use [19], European Medicines Agency (EMA) reflection paper formulations of choice for the paediatric population, 2006 [8], article by Fabiano et al. [20] and book from Costello et al. [21]. A PubMed database search was conducted by using the name of each excipient AND/OR synonyms AND "human toxicity" as search terms; no other limiters or terms were used to narrow or widen the search. If there were no results in the PubMed search or other abovementioned information sources, Google scholar (http://scholar.google.com; last accessed 24 th September 2011) search was conducted using the same search terms.
In this study, all excipients for which according to the abovementioned sources there were some safety concerns were classified as "potentially harmful" (Category 2 in Table 1).
The study was approved by the Ethics Review Committee of the University of Tartu. The study used anonymised data collected in routine clinical practice and did not require individual consent of the parent.

Classification of excipients and neonatal exposure
One third of excipients (47/123) were classified as potentially harmful (Category 2, Table 1), including eight excipients already known to be harmful in neonates (Table 2). Another third of excipients (42/123) were classified as potentially safe (Category 1, Table 1). For the remaining 34 excipients human safety / toxicity data was not found in the literature (19/34; Category 3) or the description in the SPC was too unspecific to conduct a literature search (15/34; Category 4), Table 1. Many flavouring agents (e.g. banana and strawberry flavour) and essential oils (e.g. orange essential oil) had to be classified to the latter group as their chemical entity was not described in the SPC.
Almost all treated neonates (339/348; 97%) received medicines with at least one potentially harmful excipient and as many as 88% (307/348) received at least one of the eight excipients known to be harmful in neonates ( Table 2). From the medicines prescribed, the median number of included excipients known to cause harm in neonates was two (interquartile range (IQR) 5-2; preterm neonates median of 3, range 1 to 15, IQR 4-2; term neonates median of 1, range 1 to 11, IQR 3-1).
Potentially and known to be harmful excipients in medicines used in neonates Figure 1 illustrates that approximately two thirds (73 /107, 68%) of all the medicines used in neonates contained at least one potentially harmful or known to be harmful excipient. The median number of such excipients per medicinal product was two and maximum was five (in simeticone oral suspension). The proportion of medicines containing potentially harmful excipients in preterm neonates was even higher than the general rate -77%. When preterm and term neonates were compared in terms of medicines with potentially safe excipients the percentage of such medicines was the same in both groups (22%).
The detailed characteristics of excipients and their potential safety issues are described in Table 3.
The most common excipients that are known to cause harm were propylene glycol and ethanol, both present in seven products. In relation to prescription frequency, the most common excipients known to cause harm were parabens (methyl-and propylparahydroxybenzoate) used as preservatives in parenteral gentamicin given to 57% of treated neonates. The gentamicin product also included another potentially harmful excipient, namely sodium metabisulphite (Table 4).
Simeticone oral suspension was the second most commonly prescibed medicine, given to 31% of neonates. The simeticone product contained two excipients known to cause harmsaccharin sodium and sodium benzoate, and also three other potentially harmful excipients -colloidal anhydrous silica, sorbic acid and sodium cyclamate ( Table 4).
Of excipients known to cause harm as an excipient in older age groups, but where there is no neonatal data, colloidal anhydrous silica and anhydrous sodium hydrogen phosphate were most commonly present, both in 10 medicines. Medicines containing sodium metabisulfite were prescribed most frequently in this category. Again its frequent exposure was driven by wide use of parenteral gentamicin.
Approximately two thirds of parenterally used products (29/47) contained some potentially harmful excipient. The situation was even worse for other formulations -all of the prescribed rectal, topical, inhalation, oral solutions and oral suspensions contained at least one potentially harmful excipient. The use of topical agents in neonates was rare and only 8/33 products contained excipient known to be toxic to neonates. Only one of the 19 orally administered solid drug formulations and one of the five ocular formulations were free of potentially harmful excipients. Not surprisingly, most of medicines free from potentially harmful excipients were parenteral single-dose antibacterial or antifungal formulations (13/ of all the 35 medicines without harmful excipients).
A total of 19 medicines were licensed for use in neonates (6 for preterm and all for term neonates). Approximately half of them (3/6 in preterm and 8/19 in term neonates) contained at least one potentially harmful excipient.
The amount of the excipients in the drug formulation was present in the SPCs for only two medications (metoclopramide injection solution and esomeprazole powder for injection).

Discussion
To the best of our knowledge this is the first study looking at the excipients in medicines given to an unselected group of hospitalised neonates. We describe the presence of over a hundred different excipients and demonstrate that almost all drugs used in neonates (including those licensed for neonates) contain at least one potentially harmful excipient. The safety of the majority of these excipients in neonates is not easily assessable as the information in the SPCs is scarce and published studies on the topic are rare. The fact that over a third of excipients present in medicines are known or suggested to be potentially harmful to neonates is even more worrisome.
Despite the lack of quantitative information, it could be that even if the excipient is known to be harmful, the daily intake will not exceed the toxic threshold due to the small quantities used in drug formulations. For example, from using a parenteral gentamicin product, a premature infant weighting 500 g and receiving a daily dose of 2 mg gets a maximum of 0.1mcg of parabens (methyl-and propylparahydroxybenzoate, parenteral formulations contain up to 0.75% parabens). When comparing this value to the allowed daily intake of 10 mg/kg body weight in adults it is obvious that the quantities are far below the toxic threshold. However, the fact that in neonates organs and thus the PK pathways are not fully matured may change the situation drastically [20]. Furthermore Allegaert et al. showed recently that a short duration of unintended propylene glycol administration at a median dose of 34 mg/kg over 48 hours was well tolerated by (pre)term neonates [29]. At the same time the authors stress that long-term safety of propylene glycol is still not established. We believe that the well-known toxic or potentially harmful excipients need careful safety assessment and determination of PK/PD profiles in neonates.
It has been stated that certain colouring and flavouring agents such as erythrosine (cherry-pink dye) need to be avoided in children [30]. This study also found that the SPC description of 15 excipients was too unspecific to allow a safety evaluation. Many of these excipients were flavouring agents such as banana flavour, cherry aroma, strawberry, raspberry flavour etc. Without further information in the SPC the safety assessment and critical evaluation of the administration of these agents to neonates is impossible in clinical practice. We found that the formulation of the most commonly used drug in Estonian neonates, gentamicin, contains well-known harmful excipientsparabens. However a paraben-free gentamicin product is also registered in Europe. With this our data are in line with van Riet-Nales et al. [28] who also showed that for 22% of oral liquid paediatric medicines contain potentially harmful excipients an alternative formulation lacking such excipient was available with the same active chemical entity. This indicates that health professionals have a low awareness on safety of excipients. We recommend that when compiling a formulary for the treatment of vulnerable patients such as neonates attention should also be paid to the identification of excipients in medicines.

Limitations
The most important limitation of our study is the lack of information on the exact amounts of excipients in medicines which precludes us making any conclusions on the real excipient exposure. This limitation was beyond our control because manufacturers do not disseminate this information. The other challenges are the use of a novel and non-validated classification system and restriction of the study to one country only. Also we did not collect information about dosage regimens since this would have been uninterpretable in the absence of quantitative information about the excipient content of the prescribed medicines. Another issue, possibly characteristic of other small markets, was the significant use (22%) of unlicensed medicines and thus unavailability of SPCs. In these cases the excipient content was recorded according to the package insert leaflets. Altough we appreciate that only excipients of intravenous, topical and opthalmic products and known to be toxic excipients have to be declared in the package leaflet, we assume that this will not significally affect our conclusions. Finally the study was conducted in a small country and thus these results cannot be generalised to other countries. These limitations do not undermine our findings that neonates are frequently exposed to a range of potentially harmful and known to be toxic excipients. Many of these limitations will be addressed in a recently commenced international research project, the European Study of Neonatal Excipient Exposure (ESNEE) [31], which evaluates the neonatal excipient exposure of over 20 European countries. The aims of the project are to show which excipients are used in formulations given to neonates treated in European hospitals, to conduct a review of toxicological studies on excipients and to investigate excipient kinetics during treatment with commonly prescribed neonatal medicines.

Conclusion
Hospitalised neonates commonly receive numerous excipients, several of which have not until now been discussed regarding their potential harm to neonates. There is an urgent need for the careful toxicological assessment of excipients as the information in the published literature is limited. Information about excipients should be made available to pharmacists and neonatologists to assist the selection of the most appropriate medicines for neonates. When excipients cannot be avoided, professionals should have access to quantitative and qualitative information that allows them to assess risk and monitor vulnerable patients appropriately.