2.3.2. Excipients

Some of the challenges in the development of formulations for neonates have recently been reviewed by Kogermann et al. (2017) [114]. Not surprisingly, this review focuses largely on pharmaceutical excipients since mainly liquid dosage forms are used in these vulnerable patients and these may require preservatives to limit microbial growth and/or an array of excipients to achieve a solution of the active or to formulate a palatable liquid. Given that most actives are not specifically developed for neonatal conditions, dosage forms may contain excipients deemed safe in adults or older children that have not been studied for neonates [14,15].

The immaturity of organs and physiological systems of the neonate is important for the 'handling' of excipients as well as active drug substances. Little is known or published about the acute or chronic effects of excipients in the very young and that which has been published indicates their vulnerability to toxic e ffects considered safe for older children and adults. The exposure limits for many excipients have been published, but often apply to adults and should not be applied to neonates unless specifically indicated. Non-clinical work in appropriate juvenile animal models may be required for excipients used for the first time in the very young, and there should be appropriate justification for the use of any excipient supported by studies of high quality.

Epidemiological studies have demonstrated the exposure of neonates to excipients and the variation between products marketed in di fferent countries [115]. Studies have also shown that excipient exposure can exceed current safety limits and that exposure may be from several medicines [116]. Measuring blood levels and the kinetics of excipients is possible in the very young and can be contemplated as part of clinical studies if necessary [117].

The formulator should first consider whether any excipient is required when developing a formulation for neonates. For example, it should not be necessary to include an antimicrobial preservative in a liquid product if designed for single use. Even an oral liquid medicine might be presented as a sterilised unit dose preparation without preservative excipients. Colours are generally not needed for this population and if intended for administration via enteral tubes, sweetening and taste masking may not be required. A risk-based approach [118] taking into account PK, ADME and safety data relevant to neonates may be taken in discussion of potential formulations with clinical and toxicology colleagues. For example, an excipient toxicity profile acceptable for single dose administration for a life-threatening illness may be quite di fferent from that for long-term administration for a condition for which e ffective remedies are already available. Of course, the overall cost of formulation development for neonates and the viability of the commercial product will also require consideration during this pre-formulation discussion.

The European Paediatric Formulation Initiative (EuPFI) STEP database [119] provides comprehensive information on excipients for children and is freely available. The European Commission, through the EMA, is revising the labelling requirements for excipients in authorised products. Supporting the thresholds for labelling requirements are extensive reviews of excipient properties, uses and toxicology [120]. Whilst not intended to provide direct guidance to the formulator, that for propylene glycol [121] serves as a good example of the way in which toxicity can be assessed and safe exposure limits calculated.

As with any product development, compatibility studies of any proposed excipients with the API are an important part of early product development.

### 2.3.3. A Shift towards Solid Dosage Forms?

The mainstay of oral product administration for neonates are liquid dosage forms. Despite being more common for the paediatric market, they tend to contain excipients with elevated toxicological risks in neonates compared to adults (e.g., ethanol, propylene glycol, benzyl alcohol, polysorbate, parabens, etc.) due to the ongoing organ development and incomplete maturation [122]. REF oral solid dosage form is by far dominating the adult market [123]. Although they may sometimes contain excipients with some toxicological concerns of their own (for example, some may be allergens or irritants), in general, excipients used in solid dosage forms have a more benign safety profile than some used for liquid medicines [124]. As discussed above (Section 2.2) neonates are not always able to swallow liquids let alone solid dosage forms even though there are many inappropriate products on the market such as capsules and tablets licensed from birth that will have to be manipulated for administration [125]. It is highly unlikely that this is based on clinical data and hence, would no longer be acceptable following the implementation of new regulations worldwide.

The attraction of solid dosage forms has led to the emergence since 2008 of flexible solid oral dosage form (FSOD) which has been actively promoted by the World Health Organization (WHO) to overcome challenges of ensuring access to suitable medicines for children [126].

FSOD are solid forms that do not have to be swallowed whole, such as dispersible tablets, effervescent tablets, orodispersible tablets, and sprinkle capsules. They are flexible in administration, though not necessarily in dose. They may provide a convenient dose unit that can be manipulated to provide a suitable form for delivery to a neonate.

For example, some tablets for older children can be cut accurately into two, four or even eight approximately equal pieces. Such tablets were developed for fixed-dose combination of zidovudine and lamivudine in fast-disintegrating subunits (per 5 kg of BW) [127]. These can be easily administered after dispersion in a liquid or milk.

Dispersible tablets offer more dose flexibility. However, the minimum adequate volume of dispersion linked to the reproducibly of accurate dose withdrawal needs to be considered concomitantly [128]. In theory, such a dispersion could be dosed via an enteral tube. The suspension should also of sufficiently low viscosity and be fine enough to pass through the enteral tubes if this is foreseen.

As effervescent tablets have high-sodium content, their use should be carefully considered for neonates. They are also clearly contraindicated in cases where the patient in hypernatraemic, who need to be sodium restricted. Dispersions probably need to be degassed prior administration.

Some tablets have been developed to be administered in a Therapeutic Nipple Shield to safely deliverer medications and nutrients to breastfeeding infants [129]. In one proof of concept study, formula or freeze-dried milk/drug tablets were formulated to demonstrate both reliable drug delivery to babies and the ability of milk to dissolve poorly water-soluble drugs [130,131]. Both are interesting concepts considering babies' exclusive milk-based diet. There is also some evidence that milk can mask the taste of some medicines. However, the risk of interfering with feeding by rendering the taste of the milk foul needs to be mitigated with decent palatability of the medicine itself.

Until recently, few studies have been undertaken on the acceptability of mini tablets for patients under six years of age. Recently, however, a study in 6 to 12 month-old children that also included some neonates (*n* = 151) demonstrated complete swallowing of one mini-tablet (82.2%) compared to 72.2% for a liquid syrup [132] (Figure 2). However, it is to be noted that this mini tablet was uncoated so would have dispersed if it were inhaled inadvertently rather than swallowed. This was for safety reasons but may have somewhat confounded the study. On the positive side, there was no evidence of choking. As this has been a concern for some time, it would appear that the acceptance of this dosage form has come a long way. Data on acceptability and swallowability of several hundred mini tablets in slightly older babies, infants and children (six months to six years) are awaited. In parallel orodispersible minitablets that can be dispersed in the mouth or in baby-friendly beverages to particles that are easy to swallow are proposed to reduce the risk of choking while allowing dosing via nasogastric tubes (NGTs) [133].

Other orodispersible dosage forms, such as thin polymeric film, have been studied in older age groups (0.5–6 years old) [134]. They are attractive as they overcome the need for swallowing a solid entity. However, there is no precedent of use in neonates. This is also true of micropellet formulations such as sprinkles. It is hoped that work with older babies will provide usability and safety evidence and translate in product development in the near future that could benefit neonates too.

**Figure 2.** Newborn child with uncoated mini-tablet in the cheek pouch before swallowing (with permission to use from Thabet et al. 2018 [132]).

### *2.4. Ability to Administer: Device Factors to Consider*

### 2.4.1. Accuracy of Small Volumes

Measuring small volumes (e.g., bolus IV injections, oral liquids, low rate IV infusions) with sufficient accuracy can be problematic if routinely available administration devices such as oral/enteral syringes and injection syringes are to be used [40]. Using a liquid product with a concentration designed to administer standard volumes to adults or older children may mean that volumes of 0.1 mL or less might be required for neonates [39,40,135–139]. This issue can be even more acute if clinical practice standardises on a single oral/enteral syringe design rather than using any specific device that may be provided by the manufacturer. To mitigate this issue, formulators should ensure that they develop the right product strength(s) so that there is no need to measure small volumes, especially that no volume is less than 0.1 mL and at the same time, the strength allows to cover body weights from 0.5 kg to 5 kg (a 10-fold range).

Calculating the required dose volume may be difficult, and dilution steps may add to calculation errors and be undertaken in an unsuitable environment. Decimal fractions involving hundredths of a mL can be confusing.

For injections the volume presented in a container should not be greater than ten times the dose for the smallest child and for drugs given by other routes of administration the risks of miscalculation or inaccurate measurement should be risk assessed and steps taken to reduce the risk [20].

It would be helpful if the device industry could develop administration devices that can accurately measure and reproducibly deliver very small volumes.

### 2.4.2. Enteral Tubes Administration

Neonates may require enteral feeding tubes to allow safe administration of enteral feeds, fluids and medicines. Older children with swallowing difficulties (sometimes related to administration of medicines only [140]) may have enteral feeding tubes inserted at different sites in the upper gastrointestinal tract for the long-term administration of food, fluids and medicines. The likelihood of administration via an enteral tube is one of the factors that will determine the investigations required to demonstrate that the drug preparation can be delivered effectively via the tube and without adverse effects. The flush volume required to ensure the whole dose is delivered must be carefully considered in relation to fluid requirements. If the product is likely to be administered via an enteral (e.g., nasogastric, nasojejunal) tube, issues such as viscosity of formulation (to permit flow of the product through neonatal tubes [e.g., 6FR/8FR] and avoid blockage), size of particles, adsorption to commonly used enteral tubes and interaction with common formula/breast milk should be investigated [20]. A recent Q&A has been produced by the Quality Working Party of EMA and sets out the potential problems

of administration of medicines through an enteral tube and the steps to be taken to investigate and minimise them [141].

### 2.4.3. Parenteral Catheters and Administration Sets

An important formulation development consideration is to assess the risk of interactions between certain drugs and the materials of the devices used to administer them. Neonatal catheters and cannulas can be made of a variety of di fferent materials including silicone, polyurethane, polyvinyl chloride and polyethylene [142]. Other components of the administration set may well be constructed from a di fferent polymer to that of the catheter being used. Thus, the nature of the incompatibility may be complex. For this reason, the formulator should be aware of the composition of medical devices and potential incompatibilities.

The most commonly observed interactions are chemisorption, mechanical trapping of drug molecules onto or within the device and leaching. The occurrence of an interaction depends on the chemical nature of a drug and of the device material [143], the number of binding sites on the material's surface (surface area, device dimensions) [143], the drug concentration [144], and the time of contact between the drug and the material's surface which is influenced by the flow rate and the length and diameter of any tubing. The pH of the formulation may influence binding by changing the ionisation state of the API and any binding sites on the material of construction of the administration device. It is also important to consider any physical entrapment due to sharp bends in the tubing or via in-line filters. In the latter case, the pore size and material of construction could be important. Where medications are used at low neonatal concentrations, the loss due to interactions may be relatively significant [104]. A well-studied example is insulin [143,145–147]. Insulin adsorbs to glass and plastic polymers, binding more strongly at low flow rates to polyvinyl chloride (PVC) than to polyethylene (PE) tubing [143]. The opioid analgesic fentanyl and some benzodiazepines (diazepam, clonazepam) used in neonatal population for sedation and seizure control have been shown to interact with PVC IV lines and bags whilst there was no interaction with PE [144,147].

Another issue to consider here is that of the potential risk posed by extractables and leachables from any polymeric components. This issue is not restricted to parenteral delivery but may be expected to have a significant impact in this patient group. Leachables may be an issue not only for administration sets but also for primary, secondary and tertiary packaging. Table 3 below shows the likelihood of packaging component interactions.


**Table 3.** Assessment of drug product leachables associated with pharmaceutical packaging/delivery systems (modified from "FDA/CDER/CBER risk-based approach to consideration of leachables" (USP—General chapter <1664>) [148]).

Term (and to an even greater extent preterm) neonates are at higher risk of increased sensitivity to toxicants due to their organ immaturity and altered metabolic function [149]. This should be taken into account when establishing the safety threshold of any extractables and leachables. O fficial guidance on how to do this is available [150]. This includes the calculation of 'safety factors' for neonates and children but does not cover preterm infants.

With the advent of ever more sensitive analytical techniques, the number of compounds that may need to be evaluated, albeit at trace levels, makes this aspect of product development very challenging. It can, however, have significant e ffects in clinical practice.

As an example, some drugs, such as amiodarone and etoposide and some formulation excipients, such as those used in lipid emulsions, are known as leaching promoters [104,151]. Diethylhexyl phthalate (DEHP) is a plasticiser that is commonly used in PVC polymers that may be used in administration sets [152]. Animal studies have shown that high levels of phthalates may cause a range of issues in animals, including cancers, endocrine disruption and kidney injury [153]. The US Department of Health and Human Services issued a report in 2006 in which the risks of DEHP to human reproductive and development was evaluated [154]. In the report, neonatal patients and in particular pre-term neonates were highlighted by the FDA and the European Commission as being particularly at risk of exposure to levels of DEHP where there may be toxicological concerns [155]. However, DEHP free administration sets may not always be readily available. Lala et al. have described a case where an urgen<sup>t</sup> need for an IV amiodarone infusion (which was infrequently used by their unit) was delayed several hours whilst attempts were made to source DEHP free components to adapt administration apparatus to avoid leaching of DEHP [104].

### 2.4.4. IV Polypharmacy

Patients cared for in intensive care units frequently require multiple IV infusions and present with limited vascular access. To avoid mixing di fferent drug solutions in one IV container and to reduce the contact time between these drugs Y-site, T- or multi-lumen connectors are used to separate drug administration [31,156,157]. Some examples are shown in Figure 3 below.

**Figure 3.** Di fferent types of connectors (adapted from IV Sets and Access Devices Product Catalog—B. Braun Medical Inc., e ffective August 2017).

In practice, multi-lumen and y-site connectors are often used to simultaneously deliver multiple compatible drugs through one port of entry. The use of these devices can allow for multiple drug infusions to converge and be administered through a single venous port, if compatibility studies are available. This is illustrated in Figure 4. However, the optimum geometry, e ffect of mixing, flow resistance and risk of inadvertent boluses due to drug flow through these parts have not ye<sup>t</sup> been su fficiently studied and compared.

**Figure 4.** Example of venous access and multiple drug administration devices.

### 2.4.5. Inhalation Devices

The device used is an integral part of dose delivery of respiratory medicines, and hence, devices have been discussed to a significant extent under the respiratory product development section. Metered dose inhalers (MDI) need to be assessed for the safety of any extractables and leachables, e.g., from polymeric components of valves used in MDI. Aqueous formulations present less of a risk. Nevertheless, there have been some notable innovations in the use of devices to aid drug delivery in unusual settings such as the delivery of aerosolised therapy to a sleeping infant (Figure 5) and around device development [158,159].

**Figure 5.** Photograph illustrating the method of aerosol administration to a sleeping infant showing the Respimat inhaler, InspiraChamber and SootherMask. (reproduced from [160] with permission from BMJ Publishing Group Ltd.).

### *2.5. Biopharmaceutical Considerations*

The rate and extent of absorption, distribution, metabolism and excretion of drugs in children are different from that encountered in adults with the greatest differences from the adult being observed in neonates [161]. The impact of a number of biopharmaceutics factors on product development has been discussed inter alia during the discussion about the development of products for the various routes of administration.

Rectal absorption in full-term neonates has been reported as excellent. Intrarectal pH was shown to be significantly lower in neonates (6.5) compared with infants (6.9), and illness had no effect [162]. Neonates' erratic absorption seems to be the limiting factor for wider use [36].

When medicines are administered via inhalation, breathing patterns can have an effect on the bioavailability and crying (which is more prevalent in neonates compared to adults) can reduce the amount of dose absorbed.

The skin barrier may not be fully formed in preterm infants, there are changes in the type and proportion of lipids present, and the development of ancillary skin structures such as sweat ducts and hair follicles differs to that of the skin of a term neonate. Skin may also be thin, fragile, hydrated and prone to occlusion. This has implications for potential toxicity of APIs and formulations usually delivered for a localised effect via the topical route.

Altered or reduced distribution is likely via most parenteral routes.

A recent review describes these differences in detail [13], the most relevant information has been extracted and is presented in Table 4.

**Table 4.** Summary of differences between neonatal and adult physiology that affect absorption/distribution of drugs (extracted from [13]).


The gastrointestinal environment is very different in neonates compared to adults due to the ontogeny of transporters, pH and permeability of the intestinal wall [6,13,163–165]. Furthermore, the impact of feeding on drug absorption needs to be considered due to the very different feeding pattern in neonates as well as the co-administration of medicines with foods to improve acceptance and palatability [166,167]. In neonates and infants, attention has been paid to the mixing of medicines with liquid feeds due to the impact on the osmolality of the resulting fluid which may affect GI transit [168,169] although the actual physiological impact is still unclear [170].

Many medicines are mixed with food to improve palatability, ye<sup>t</sup> the impact of this mixing on the absorption of the drug is unknown [171]. The dissolution of drug products in breast and formula milk has been evaluated in vitro to identify whether any effects can be predicted [172]. Other paediatric-relevant dissolution testing systems have also been introduced [173] as well as media that represents the gastric and intestinal fluid from children [163].

Doses used in paediatric populations are often derived from weight-based and surface-area-based dosing equations that are often based on adult data and then scaled based on size and age as an approximation for drug activity in children. However, paediatric growth and development is not a linear process, and there are risks associated with simple scaling to determine doses to use in neonates.

Comprehensive physiologically based pharmacokinetic modelling (PBPK) systems have been introduced that replicate the known parameters of the paediatric GI tract to better predict oral dosing [174–182]. However, the major limitation in both in vitro and in silico models is the lack of accurate knowledge about the neonatal and infant intestinal parameters [178,183,184]. These models have focused on oral absorption to date, and there is a need to develop appropriate tools for other routes of drug administration in neonates.

Clinical examples of di fferences in absorption in neonates compared to adults are scarce. Recent work suggested that the processes underlying changes in oral drug absorption rate typically reach adult levels within one week of birth [35].

### *2.6. Regulatory Challenges*

Useful background information on the European Paediatric Regulation, together with progress reports, is published by the European Commission [7,185]. The US FDA has a similar regulatory framework and many other countries model their approach to the assessment of products on these two sets of regulations [186]. The clarity that these regulations and associated guidelines have provided has advanced the number of new medicines that are available for children. However, the 2017 review of 10 years of the implementation of the European regulations shows there is still a long way to go [185].

Many of the medicines used for neonates have only been developed and authorised for adults and older children. That is, they are used o ff-label. They may sometimes have been studied extensively perhaps 'in-house', but information has often been gathered and published by academia and never submitted by the pharmaceutical industry to regulatory scrutiny [187,188]. Frequently the dosage form that is authorised is not acceptable for neonates or younger children and must be modified for administration via an extemporaneous preparation by the pharmacist or manipulation at the point of administration such as splitting or crushing a solid oral dosage form [189]. Modification of the dosage form may generate an 'unlicensed' medicine with attendant uncertainty and risks. Although there is a strong drive to develop licensed medicines for all paediatric subpopulations, as noted throughout this paper this can be extremely challenging for neonates and in some cases industry verified manipulations can provide some increased level of assurance of suitability for use in the interim before a suitable neonate formulation can be licensed [188].

Whilst regulation demands a paediatric investigation plan (PIP) for all new medicines, this does not apply to long-established molecules. There are incentives to study and seek authorisation for age-appropriate dosage forms of o ff-patent medicines used by children (the Paediatric-Use Marketing Authorisation (PUMA) [190]), but it has not been considered successful in providing the medicines required. The incentives do not appear to be enough to balance the cost of development studies, and despite the acknowledged benefits of using a licensed product, the inevitable cost of such medicines can limit market entry. Enforcement of the use of licensed medicines appears not to be systematic and unlicensed medicines continue to be used.

Since it appears that extemporaneous modification is likely to persist the European Directorate for the Quality of Medicines (EDQM) is establishing a formulary of paediatric medicines prepared extemporaneously. As of January 2019, two monographs are available for public consultation. This seeks to improve access to medicines of demonstrable quality when suitable marketed products are not available [191].

### **3. Burden of Proof**

Given all the product, environmental, biopharmaceutics, dose delivery and device factors discussed above, along with the multiplicity of di fferent practices in the clinical setting based upon the age of the neonate, their stage of development and their clinical need, the potential number of pharmaceutical development studies and in particular the compatibility and stability studies are huge. There is a complex matrix of confounding factors such as polypharmacy/coadministrations, low or high concentrations, long residence time and high risk of adsorption to the medical devices

used for the administration, e.g., tubing, extractables and leachables assessment, interaction with packaging and unique environmental conditions many of which either alone or in combination can lead to unfavourable clinical outcomes.

The principal aim of a formulator is to design a formulation that provides the intended e fficacy for the target patients, ensuring a safe and robust quality profile for the final drug product. However, it would perhaps be unrealistic to test every possible combination. Although this is part of current practices of drug product development, it is important in the context of neonates to reinstate that product developers provide evidence and advice that will cover the majority of anticipated use cases covering both what is acceptable and what is contraindicated from a product quality, safety and e fficacy point of view. A lot of this advice can be extrapolated from studies performed for other paediatric subpopulations or adults. However, specific pharmaceutical development studies taking into account the specific requirements of neonates may well be required.

If clinical practice demands that manipulation of formulations or co-dosing etc. that is outside the range of studies performed by the product developer or reported and justified in the literature, then it is the responsibility of the team treating that patient to consider the wisdom of the proposed action. Clearly the pharmacist supporting that team can be of significant importance in evaluating the product quality aspects. The level of practical work supporting those decisions will depend on the circumstances. Thus a 'one o ff' intervention may perhaps be supported on a risk/benefit judgement based on general medical and pharmaceutical principles. If the intervention is used more frequently, then the pharmacy should undertake and publish appropriate compatibility and stability studies. If the intervention is to become commonplace, then formal studies perhaps in association with the original product developer and/or academia will be warranted.

It is essential to assign to appropriate shelf life and/or in-use period for the final product under specific storage conditions (i.e., temperature, humidity and light). Conditions in the NICU may, in fact, represent the best case-controlled environment for the delivery of medications to neonate and drug formulators should investigate other environments in which uncontrolled environmental conditions may present challenges not seen in the NICU. It is vital to demonstrate the in-use usability of the entire product including storage requirements in the likely use environment, access to the packaging, reconstitution procedure, dose measurement, suitability and ease of use of any device required, clarity of instructions. All these factors can/should be assessed in a simulated use evaluation supervised by clinical and/or pharmacy sta ff using a representative sample of professional sta ff (i.e., nurse, physician, pharmacy technicians) and where relevant non-professional carers such as parents and guardians. With the increased involvement of sta ff and carers in the treatment of neonates, evaluation of the role of human factors could be particularly relevant during development.

Studies should aim to mirror clinical practice where possible bearing in mind that the needs for each neonate could vary significantly and might require customized treatment regimens, especially as new treatments become available. Usability study should also include the stability of the formulation over the extended "worst case" time periods being cognisant of regulatory policies on microbiological stability. Limited venous access in neonates can in clinical practice result in little choice but to co-administer drugs. Formulators should be aware of this potential and should extend stability studies to examine Y-site stability of co-administered drugs and also TPN [109,157].
