Drug dose estimation without pharmacokinetic data

The dose regimens of many drugs administered to children have historically been based on pharmacokinetic data generated in adults. The adult dose was simply adjusted to the child's body weight. However, it is clear that on-going developmental changes can significantly alter drug pharmacokinetics.6 Consequently, dosing children on the basis of data generated in adults (corrected for the child's body surface area or weight) is more likely to lead to overdosing and toxicity. Chloramphenicol and propofol are important examples of children being prescribed inappropriately high and toxic doses of medicines because of unrecognised, age-specific differences in pharmacokinetics.7 8 There are few instances of drugs prescribed at subtherapeutic doses based on adult pharmacokinetic data. This may reflect the difficulty in detecting treatment failure due to age-related differences in pharmacokinetics. Aminophylline is an example of a drug that is required to be prescribed in higher doses (per mg/kg) in lower age groups to achieve the same therapeutic level.9 The overall effect is that potentially useful drugs are perceived as either harmful or ineffective when the real issue may be one of a lack of pharmacokinetic data in the relevant age group.

The lack of pharmacokinetic data in paediatric medicine has also led to a ‘trial-and-error’ approach to finding an optimal dosing regimen. A high profile example is dexamethasone therapy for chronic lung disease (CLD) of prematurity.10 Only two studies involving 16 patients in total have investigated the pharmacokinetics of dexamethasone.11 12 The data from these studies are insufficient to confidently derive the plasma clearance and volume of distribution of dexamethasone. Without these pharmacokinetic parameters, it is not possible to estimate the optimal dexamethasone dose for infants between 24 and 36 weeks' gestation. The paucity of dexamethasone pharmacokinetic data for this group has led to initial trials with ‘high’ dexamethasone doses for prolonged periods with more recent trials using ‘lower’ doses for a shorter period. Despite several trials, incorporating various dosing schedules, the optimum dexamethasone dosing regimen remains unknown. Currently, some neonatologists do not use dexamethasone, whereas those who do cannot be certain of the optimal dosing regimen. An initial study exploring the pharmacokinetic–pharmacodynamic relationship of dexamethasone would provide an efficacious dose range around which to design a much more efficient and possibly more conclusive trial. Similar criticisms could be applied to the current national trial of nebulised magnesium in acute asthma.13 If the drug shows no or minimal efficacy in this trial, it may not be possible to judge whether it has a role in the management of childhood asthma given the paucity of pharmacokinetic–pharmacodynamic data. In contrast, clinical efficacy trials for montelukast and caffeine based on pharmacokinetic–pharmacodynamic data are examples of well-founded paediatric drug studies.14 15

Blood sampling: an obstacle to performing pharmacokinetic studies in children

The problem of repeated blood sampling has been obviated through the use of population pharmacokinetic study designs that incorporate sparse sampling methodology. This means that only two to three blood samples are needed from each child in a study rather than in excess of 12 samples required by traditional pharmacokinetic modelling methods. The population pharmacokinetic approach is able to reliably estimate the values of pharmacokinetic parameters such as clearance, volume of distribution and half-life and their associated variability within the population. Indeed, the population pharmacokinetic approach has been recommended by the regulatory agencies as the preferred approach to studying pharmacokinetics in children.16

However, an additional, substantial barrier to conducting pharmacokinetic trials in children is the relatively large volumes of blood (∼1–10 ml at a time) required by most drug assays. This issue is most apparent in preterm neonates because of their comparatively small circulatory volume. Draft guidelines on obtaining blood from neonates for research purposes recommend that “trial-related blood loss should not exceed 3% of the total blood volume during a period of four weeks and should not exceed 1% at any single time”.17 This equates to 1.2 ml over 4 weeks for a baby weighing 500 g.

In circumstances where the volume of blood does not prevent a pharmacokinetic study from taking place, blood sampling can make children, parents and ethics committees baulk at agreeing to a pharmacokinetic study. Moreover, all blood sampling methods have a failure rate depending on the volume of blood needed and the skill of the phlebotomist. A capillary stab nearly always results in drawing a small volume of blood and is technically less demanding than venepuncture. However, it is often not possible to collect large volumes of blood using this method. For these reasons, it is clear that pharmacokinetic data on drugs used in all age groups will only become readily available through the development of assays and methods that are compatible with micro-volume (≤50 µl) capillary blood samples. Moreover, these new methods must be acceptable to parents and children and validated so that everyone, including medicines regulatory agencies, have confidence in the data.

Factors of importance in DBS analysis

The DBS technique as applied to pharmacokinetic studies is different to that applied for newborn screening programmes. The blood is collected into a capillary tube (containing anticoagulant) and then spotted onto the card (filter paper) using a suction bulb (figure 1). The latter ensures complete emptying of the tube of blood. Moreover, this method reduces the risk of the card being contaminated or damaged through contact with the capillary tube. It also facilitates an even spreading of blood across the sample collection area. It is essential that the same volume of blood occupies a consistent area of the DBS card. This is because a disc punched from a DBS card provides a volumetric measurement akin to the use of a pipette for liquid measurements.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Controlled application of blood onto a filter card using a pipetting device. A disc of blood is subsequently cut from the dried blood spot for analysis (left). Capillary tube with suction bulb attached (right).

Several groups have found that haematocrit and the volume of blood applied onto a filter card are the two most influential variables affecting drug measurements using DBS sampling.32 Haematocrit has a strong influence on blood viscosity and hence spreading of blood when applied to a card. A high haematocrit value reduces blood spreading, which in turn can reduce the size of the blood spot. This results in a higher concentration of drug per unit area of filter card. However, depending on the drug, moderate variations in haematocrit may be acceptable. For example, haematocrit values between 30% and 55% do not significantly alter caffeine concentrations measured by the DBS method. In addition, for some DBS-based drug assays, it may be possible to introduce a correction factor to account for the effect of haematocrit. Applying too much or too little blood can also affect drug concentration measurements using DBS. The reasons for this are incompletely defined but include uneven distribution of plasma over the filter card. As with haematocrit, there is a margin within which differences in the volume of blood spotted has little effect on assay accuracy.29

Clearly, as with traditional plasma and whole blood assays, it will be necessary to detail the assay methods used to measure a particular drug by the DBS technique. This will include factors such as the type of anticoagulant, the blood volume, the filter paper type and the drying and storage methods for the DBS card.

Barriers and challenges to implementing DBS for paediatric pharmacokinetic studies

Training issues

Children in hospital settings will form a large proportion of the target population for pharmacokinetic studies as blood sampling expertise is currently concentrated in hospitals. While the procedure is simple, research staff will need to be trained in the methods. However, this can be easily overcome by providing visual and written training materials (eg, videos, pamphlets) and face-to-face training sessions.

DBS methods are accepted by drug regulatory agencies such as the European Medicines Agency for licensing purposes. The paediatric community will need to be aware of the differences between drug measurements as quoted using plasma and DBS (whole blood) assays. This issue may be particularly important for drugs that require therapeutic monitoring.

Fundingpharmacokineticstudies

It is mandatory for pharmaceutical companies to conduct appropriate pharmacokinetic studies for drugs in clinical development. However, they are unlikely to investigate off-patent medicines currently used on an unlicensed or off-label basis, such as dexamethasone for CLD and nebulised magnesium for asthma. For these drugs, the responsibility for performing pharmacokinetic studies is likely to fall on national government funding agencies (eg, medical research councils in Europe, the US National Institutes of Health), the EU and charities such as the Wellcome Trust.

Parent/staffacceptability

DBS technology provides a minimally invasive method for drug quantification. However, the methodology will only be useful for pharmacokinetic studies if it is well accepted by parents and clinical staff.