Review
Approaches to Intravenous Clinical Pharmacokinetics: Recent Developments With Isotopic Microtracers
The Journal of Clinical Pharmacology 2016, 56(1) 11–23 © 2015, The American College of Clinical Pharmacology DOI: 10.1002/jcph.569
Graham Lappin, PhD
Abstract Obtaining pharmacokinetic data from the intravenous route for drugs intended for oral administration has traditionally been expensive and time consuming because of the toxicology requirements and challenges in intravenous formulations. Such studies are necessary, however, particularly when regulator agencies request absolute bioavailability data. A method has emerged whereby the drug administered intravenously is isotopically labeled and dosed at a maximum of 100 mg concomitantly with an oral administration given at a therapeutically relevant level. The intravenous administration has been termed a microtracer and obviates intravenous toxicology requirements as well as simplifying formulations. The study design also essentially removes issues of nonlinear pharmacokinetics that may occur when oral and intravenous doses are administered separately. This review examines the methodology and the literature to date, including those studies intended for regulatory submission. The method has been extended to the study of prodrug–to–active drug kinetics and to obtaining clearance, volume of distribution, and absolute bioavailability at steady-state conditions.
Keywords isotopic tracer, microtracer, absolute bioavailability, accelerator mass spectrometry, clearance, volume of distribution, intravenous administration, steady state
It has been estimated that approximately 90% of smallmolecule drugs (excluding anti-infectives) are developed primarily for oral dosing,1 as this route offers the most convenient and noninvasive method of administration. Although for many drugs, intravenous dosing is not relevant to clinical practice, pharmacokinetic (PK) data from this route of administration is nevertheless extremely valuable, as this is the only way of determining a drug’s fundamental PK in humans. Along with clearance (CL), volume of distribution (V), and fraction of the administered dose reaching systemic circulation (F, otherwise known as absolute bioavailability), intravenous administration allows for the deconvolution of absorption and dissolution rates, absorption rate kinetics, and potential flip-flop kinetics to be studied.2 From a regulatory perspective, there are no specific requirements to determine CL or V, but the regulatory agencies in Europe, the United States, and Japan may request absolute bioavailability data when bioavailability is apparently low or variable, and there is a known relationship between pharmacokinetics and pharmacodynamics.3 The Australian regulator, however, has required absolute bioavailability data as part of their registration package since 2006.4 Nevertheless, a case can be presented to regulators that a particular regulatory requirement is unnecessary or impractical, and absolute bioavailability is no exception (a so-called biowaiver application, discussed below). Irrespective of the regulatory requirements for absolute bioavailability, intravenous
PK data are useful to a drug developer because the steady-state plasma concentrations attained after repeat dosing are related to CL. Changes in blood flow due to age or disease may alter CL which then affects drug safety and dosing regimens. Volume of distribution can be subdivided into Vc (volume of the central compartment), Vss (volume of distribution at steady state), and Vz (volume of distribution during the elimination phase). Each parameter has its own applications, but Vss is generally considered the most calculated parameter, as it can be estimated from noncompartmental methods. It is particularly useful when the drug is intended for repeat administration, as it represents the apparent distribution volume associated with the plasma drug concentration in steady-state conditions.5 Moreover, V and CL are important, as the combination of these parameters drives half-life (t1/2), which is important when assessing the duration of therapeutic
Visiting Professor of Pharmacology, School of Pharmacy, University of Lincoln, Lincoln, UK Submitted for publication 30 April 2015; accepted 8 June 2015. Corresponding Author: Graham Lappin, PhD, Visiting Professor of Pharmacology, School of Pharmacy, University of Lincoln, Joseph Banks Laboratories, Green Lane Lincoln, LN6 7DL, UK Email: [email protected]
The Journal of Clinical Pharmacology / Vol 56 No 1 2016
12 action following a single dose or the time taken to achieve steady-state kinetics following repeat dosing. The overall pivotal importance of V and CL in drug development has been known for many years but largely overlooked, probably because of the challenges of conducting intravenous studies in humans, as discussed below.6 When a drug is not intended for intravenous administration in clinical practice, there is an understandable reluctance to develop an injectable form just for the study of its fundamental PK. In the past the industry largely attempted to address the requirement for bioavailability data by combining information from a number of studies and knowledge of a drug’s Biopharmaceuticals Classification System (BCS), although this was sometimes insufficient and an intravenous study was still necessary.7 Traditionally, an intravenous clinical PK study was expensive and time consuming, as it required toxicological testing via the intravenous route in 2 species (rodent and nonrodent). In addition, the drug must be formulated so that it will not precipitate following injection, and this can be challenging for poorly watersoluble compounds. Ezetimibe, for example, is a BCS class II, orally administered drug with relatively recent marketing authorization (May 2013) that has not been administered intravenously to humans because it is insufficiently water soluble.8 Being a BCS class II drug, however, ezetimibe has high permeability and is therefore probably well absorbed, and yet its absolute bioavailability is unknown. It must also be borne in mind that in some cases in which lipophilic drugs are successfully formulated for intravenous administration, the formulation itself may then influence its pharmacokinetics.9 Some estimates have put the costs of obtaining regulatory authorization for an intravenous PK study at $2 million and taking up to a year.7,10 This cost and effort may then only apply to conduction of a single clinical study to provide the appropriate PK data. Companies have therefore tended to perform clinical PK studies using the intravenous route (for drugs not otherwise intended for vascular administration) very reluctantly and only when regulatory authorities have demanded them.
Traditional Approaches to Intravenous Clinical Pharmacokinetic Studies The cost and resource challenges of intravenous administration to humans has traditionally limited the conduct of these types of study to situations in which a regulatory agency has asked for clinical absolute bioavailability data. A typical study involves a 2-way crossover design in which a single cohort of subjects is administered the drug extravascularly (typically oral) on 1 dosing occasion and intravenously on the other dosing occasion (separated by a suitable washout period). Diet, environment, and the consumption of other medications are controlled as far as
possible to minimize variables across dosings (ie temporal effects). The routes of administration may also be randomized between dosing occasions so that temporal effects can be minimized. Of course, it is not possible to entirely control these aspects, and the introduction of variables between dosing occasions that might affect the data cannot be ruled out entirely. Following dosing, the plasma–drug concentrationversus-time data are acquired for both dose routes, and the absolute bioavailability (F) is calculated by comparing the area under the drug concentration–time curve extrapolated to time-infinity (AUC0–1) for the extravascular (AUCex) and intravenous (AUCiv) doses, normalized to the doses administered (equation 1). F¼
AUC ex AUV iv
Doseiv Doseex
ðEquation 1Þ
Although equation 1 is commonly used in the determination of F, it should be remembered that it does not show the influence of CL. The 2 equations in respect to CL are shown in equations 2 and 3 below. CLiv ¼
Doseiv AUC iv
ðEquation 2Þ
CLex ¼
ðF Doseex Þ AUC ex
ðEquation 3Þ
Thus, F calculated from equation 1 assumes that CL is the same for both oral and intravenous administration. This is probably a reasonable assumption providing CL is not plasma drug–concentration dependent. The ideal crossover study design, therefore, is one in which the plasma concentrations (and hence AUCs) attained for the oral and intravenous doses are the same, or at least as close as possible. In practice, it is very difficult to design a study that will reliably achieve equivalent AUCs across the dose routes without already knowing F and V. There are cases, for example, in which apparent absolute bioavailability appears to be in excess of 100% because of membrane transporter–mediated nonlinear PK.11 Nonequivalent clearance is also thought to be responsible for an approximate 10% error in the determination of F for phenytoin.12 Error in the calculation of F arising from nonequivalent CL is rarely considered in absolute bioavailability studies, although the effect has been recognized for several decades. The crossover absolute bioavailability study design is limited to single-dose pharmacokinetics and does not take into account changes in V, CL, and F that might arise after repeat dosing. For chronically administered drugs in particular, it might be more appropriate to measure V, CL, and F under steady-state conditions. Calculation of V, CL, and F at steady state has been attempted by various
Lappin
13
Figure 1. The general principle behind the concomitant oral and intravenous isotopic tracer method to determine absolute bioavailability (F) and other PK parameters such as CL and V. The specific applications are explained in the text.
mathematical methods based on single-dose data. Vss, for example, is commonly calculated from the product of CL and mean residence time. Steady state conditions via the intravenous route can also be simulated for a limited time by a slow infusion. A method that experimentally measured these PK parameters under long-term steadystate conditions would be preferable, however.
Advent of Intravenous Clinical Pharmacokinetic Studies Involving Isotopic Tracers The issue of nonequivalent CL in the determination of F was recognized in the late 1960s to mid-1970s, which led to an alternative approach to the crossover design.13,14 Instead of administering the oral and intravenous doses on 2 separate dosing occasions, both doses were administered simultaneously, but the drug given intravenously was labeled with an isotopic tracer. In the first clinical study of this type by Strong et al in 1975,14 the absolute oral bioavailability (along with CL, V, and CLR) of N-acetylprocainamide (NAPA) was determined by administering a 500-mg single oral dose simultaneously with an intravenous dose of 250 mg 13C-NAPA (although the original study by Strong et al administered the oral and intravenous doses simultaneously, subsequent studies had a period of an hour or more between dose administrations, and so it is more correct to call the dosing concomitant;
this is explained further below). The plasma drug concentrations were independently determined using gas chromatography–mass spectometry for nonlabeled NAPA (from the oral dose) and 13C-NAPA (from the intravenous dose) by means of their different molecular masses. Using this method, both the oral and intravenous PK were determined from the drug concentration–time curves obtained from the analysis of a single set of plasma samples taken over time following the 1 dosing occasion. Because the overall systemic drug concentration was equal to the sum of that arising from the oral and intravenous doses, questions of nonequivalent kinetics that might otherwise occur between separate dosing occasions were virtually eliminated. In addition, this study design with its single dosing occasion eliminated temporal effects between dosings. The general principle of the isotopic tracer technique is illustrated in Figure 1. The study design employed by Strong et al had a particular shortcoming in that the intravenous administration of 250 mg made a significant contribution to the systemic drug concentration in addition to that delivered by the oral dose of 500 mg, which then had to be accounted for in the PK calculations (a multicompartmental model was used in this particular case). A better design would have been to administer a very small amount of drug by the intravenous route so that it did not significantly add to the plasma drug concentration derived from the oral dose. If the intravenous dose was kept sufficiently low, then it
14 would be reasonable to assume that the total plasma drug concentrations arose from the oral dose alone, thereby removing the need to deconvolute the kinetics. In the mid1970s, however, the sensitivity of analytical technology limited the degree to which the intravenous isotopically labeled dose could be lowered. The technique of Strong et al was adopted in a number of studies that followed utilizing a variety of isotopic tracers including 13C,15–19 15N,20,21 14C,22 18O,23 and 2 24 H. In addition, other approaches to estimating F were developed, including semisimultaneous dosing, in which an intravenous dose was given shortly after an oral dose (both nonlabeled) and the AUCs deconvoluted by mathematical modeling (known as Monte Carlo simulation).25,26 Alternative modeling methods attempted to estimate bioavailability from oral data alone.27,28 None of these methods were widely adopted, however, because of the inherent uncertainties of modeling compared with those based on an empirical approach.27,29 Moreover, there are uncertainties whether the regulatory authorities would accept such modeling studies as definitive data in a regulatory submission. The isotopically labeled intravenous dose approach first developed by Strong et al addressed potential issues of nonequivalent kinetics and removed questions of temporal effects between dose occasions, but the cost and effort surrounding regulatory approval for the intravenous administration remained, and so clinical absolute bioavailability studies still remained unpopular with the industry. However, the situation changed in 2006.
Advances in Isotopic Tracer Methods for Clinical Pharmacokinetics In 2006 the Australian Therapeutic Goods Administration Pharmaceutical Subcommittee announced that absolute bioavailability studies would be a requirement for all new chemical entities.4,7 As stated above, regulatory agencies might accept an argument that a particular study is unnecessary or impractical if a sound case can be made. The Australian agency’s announcement, however, made the exclusion of an absolute bioavailability study more difficult, and many companies were resolved to conducting an absolute bioavailability study as part of the registration package.7 It was also coincidently in 2006 that a publication appeared that reported the absolute bioavailability of 3 drugs using the isotopically labeled intravenous dose method (as depicted in Figure 1), in which the intravenous administration was only 100 mg each.30 For reasons that will become apparent, the isotope used in these investigations was 14C. This very low intravenous dose did not significantly contribute to systemic drug concentrations arising from the oral dose (oral doses ranged from 7.5 to 250 mg depending on the drug; see below) but demanded a
The Journal of Clinical Pharmacology / Vol 56 No 1 2016
very sensitive analytical method to determine the plasma 14 C-drug concentrations. The chosen method of 14C-drug analysis was accelerator mass spectrometry (AMS)31 that under the appropriate conditions can measure 14C in the low attomole range.32 The 3 drugs were erythromycin, midazolam, and ZK253 (a drug dropped from the Schering development pipeline because of poor oral bioavailability). In each case the drug (erythromycin, midazolam, or ZK253) was administered orally at a therapeutic dose level concomitantly with a 14C–isotopically labeled 100-mg intravenous dose. Plasma samples were analyzed with both liquid chromatography–tandem mass spectrometry (LC-MS/MS; tuned to the 12C ion) and by high-pressure liquid chromatography (HPLC) separation of parent drug followed by isotopic 14C AMS analysis. Using AMS enabled drug concentrations as low as 100 fg/mL plasma to be quantified. Around this time the regulatory agencies started to publish guidelines on the safety toxicology requirements necessary to allow the administration of very low doses of drug, termed a microdose (defined as 1% the pharmacologic dose or 100 mg, whichever is the lower). Microdoses are used in so-called phase 0 studies, principally to attain preliminary pharmacokinetic data in humans prior to phase 1. Because the toxicological risks associated with such low doses are inherently low, the regulatory guidelines allowed administration of a maximum of 100 mg on the basis of a much reduced safety toxicology package. Although a microdose and the intravenous administration of an isotopically labeled drug discussed in the current review are different concepts, the regulatory guidelines treated them very similarly. In fact, the guidelines called both types of administration a microdose, which has led to some considerable confusion and misinterpretations of data in the literature (examples are discussed in Lappin et al, 201333). Because of this, a new term has started to appear in the literature that refers to the intravenous dose of an isotopically labeled drug in an absolute bioavailability study administered as 1% of the pharmacologic dose or 100 mg, whichever is the lower, as a microtracer.34 The use of the term microtracer distinguishes the method discussed in this review from a microdose and hopefully will avoid future confusion in the terminology. This review uses the term microtracer henceforth. The most recent regulatory guideline, ICH M3, enables a microtracer to be administered intravenously concomitantly with a therapeutic oral dose without any additional safety data other than that required for the therapeutic oral dose plus corresponding systemic exposure data.35 Thus, by reducing the intravenous dose to a maximum of 100 mg, the isotopic tracer design not only essentially removed issues of nonequivalent kinetics and temporal effects but also eliminated the expense of conducting intravenous toxicology. In addition, the formulation of
Lappin
15
Table 1a. Summary of Microtracer Studies Published in the Peer-Reviewed Literature to Date Intended for Research or Method Development PK Parameters Determineda
Drug Erythromycin, Midazolam, ZK253 b-sitosterol R-142086 (Daiichi-Sankyo) Fexofenadine Clarithromycin Sumatriptan Propafenone
Isotope
i.v. Dose
CL, V, F
14
100 mg
CL, V, Vss, F, metabolic turnover CL, V, F
14
3-4 mg
14
100 mg
CL, V, Vss F CL, V, Vss, F
14
100 mg 100 mg
C
C C C C
14
Notes
Reference Lappin et al (2006)30
Microtracer over a dietary intake of b-sitosterol In dogs. Microtracer compared to cross-over
Duchateau et al (2012)43 Miyaji et al (2009)49 Lappin et al (2010)50 Lappin et al (2011)51
A microtracer is defined as an isotopically labeled intravenous administration of 100 mg or less given concomitantly with an extravascular dose. To the author’s knowledge, in all cases in which 14C drug and AMS were used, the levels of administered radioactivity were below those requiring regulatory approval, and therefore no dosimetry studies were necessary. Also to the author’s knowledge, intravenous doses were administered without additional intravenous toxicology data other than possibly an occasional local tolerability study for the intravenous injection site. a PK parameters relevant to the microtracer administration excluding t1/2 (and by implication Kel) and C0, which were obtained in all cases.
100 mg, or less of a drug for intravenous dosing is significantly less problematic than with higher concentrations. Indeed, some studies (indicated in Tables 1a and 1b) reduced the intravenous dose to much below 100 mg when the necessity arose. Formulation of such small intravenous doses, even with relatively insoluble drugs, can usually be achieved in vehicles such as physiological saline or glucose. Taking this approach also avoids risks of PK artifacts that might be caused by more esoteric dose
formulations.9 In cases of highly insoluble drugs, the dose can be lowered (eg,
Approaches to Intravenous Clinical Pharmacokinetics: Recent Developments With Isotopic Microtracers
The Journal of Clinical Pharmacology 2016, 56(1) 11–23 © 2015, The American College of Clinical Pharmacology DOI: 10.1002/jcph.569
Graham Lappin, PhD
Abstract Obtaining pharmacokinetic data from the intravenous route for drugs intended for oral administration has traditionally been expensive and time consuming because of the toxicology requirements and challenges in intravenous formulations. Such studies are necessary, however, particularly when regulator agencies request absolute bioavailability data. A method has emerged whereby the drug administered intravenously is isotopically labeled and dosed at a maximum of 100 mg concomitantly with an oral administration given at a therapeutically relevant level. The intravenous administration has been termed a microtracer and obviates intravenous toxicology requirements as well as simplifying formulations. The study design also essentially removes issues of nonlinear pharmacokinetics that may occur when oral and intravenous doses are administered separately. This review examines the methodology and the literature to date, including those studies intended for regulatory submission. The method has been extended to the study of prodrug–to–active drug kinetics and to obtaining clearance, volume of distribution, and absolute bioavailability at steady-state conditions.
Keywords isotopic tracer, microtracer, absolute bioavailability, accelerator mass spectrometry, clearance, volume of distribution, intravenous administration, steady state
It has been estimated that approximately 90% of smallmolecule drugs (excluding anti-infectives) are developed primarily for oral dosing,1 as this route offers the most convenient and noninvasive method of administration. Although for many drugs, intravenous dosing is not relevant to clinical practice, pharmacokinetic (PK) data from this route of administration is nevertheless extremely valuable, as this is the only way of determining a drug’s fundamental PK in humans. Along with clearance (CL), volume of distribution (V), and fraction of the administered dose reaching systemic circulation (F, otherwise known as absolute bioavailability), intravenous administration allows for the deconvolution of absorption and dissolution rates, absorption rate kinetics, and potential flip-flop kinetics to be studied.2 From a regulatory perspective, there are no specific requirements to determine CL or V, but the regulatory agencies in Europe, the United States, and Japan may request absolute bioavailability data when bioavailability is apparently low or variable, and there is a known relationship between pharmacokinetics and pharmacodynamics.3 The Australian regulator, however, has required absolute bioavailability data as part of their registration package since 2006.4 Nevertheless, a case can be presented to regulators that a particular regulatory requirement is unnecessary or impractical, and absolute bioavailability is no exception (a so-called biowaiver application, discussed below). Irrespective of the regulatory requirements for absolute bioavailability, intravenous
PK data are useful to a drug developer because the steady-state plasma concentrations attained after repeat dosing are related to CL. Changes in blood flow due to age or disease may alter CL which then affects drug safety and dosing regimens. Volume of distribution can be subdivided into Vc (volume of the central compartment), Vss (volume of distribution at steady state), and Vz (volume of distribution during the elimination phase). Each parameter has its own applications, but Vss is generally considered the most calculated parameter, as it can be estimated from noncompartmental methods. It is particularly useful when the drug is intended for repeat administration, as it represents the apparent distribution volume associated with the plasma drug concentration in steady-state conditions.5 Moreover, V and CL are important, as the combination of these parameters drives half-life (t1/2), which is important when assessing the duration of therapeutic
Visiting Professor of Pharmacology, School of Pharmacy, University of Lincoln, Lincoln, UK Submitted for publication 30 April 2015; accepted 8 June 2015. Corresponding Author: Graham Lappin, PhD, Visiting Professor of Pharmacology, School of Pharmacy, University of Lincoln, Joseph Banks Laboratories, Green Lane Lincoln, LN6 7DL, UK Email: [email protected]
The Journal of Clinical Pharmacology / Vol 56 No 1 2016
12 action following a single dose or the time taken to achieve steady-state kinetics following repeat dosing. The overall pivotal importance of V and CL in drug development has been known for many years but largely overlooked, probably because of the challenges of conducting intravenous studies in humans, as discussed below.6 When a drug is not intended for intravenous administration in clinical practice, there is an understandable reluctance to develop an injectable form just for the study of its fundamental PK. In the past the industry largely attempted to address the requirement for bioavailability data by combining information from a number of studies and knowledge of a drug’s Biopharmaceuticals Classification System (BCS), although this was sometimes insufficient and an intravenous study was still necessary.7 Traditionally, an intravenous clinical PK study was expensive and time consuming, as it required toxicological testing via the intravenous route in 2 species (rodent and nonrodent). In addition, the drug must be formulated so that it will not precipitate following injection, and this can be challenging for poorly watersoluble compounds. Ezetimibe, for example, is a BCS class II, orally administered drug with relatively recent marketing authorization (May 2013) that has not been administered intravenously to humans because it is insufficiently water soluble.8 Being a BCS class II drug, however, ezetimibe has high permeability and is therefore probably well absorbed, and yet its absolute bioavailability is unknown. It must also be borne in mind that in some cases in which lipophilic drugs are successfully formulated for intravenous administration, the formulation itself may then influence its pharmacokinetics.9 Some estimates have put the costs of obtaining regulatory authorization for an intravenous PK study at $2 million and taking up to a year.7,10 This cost and effort may then only apply to conduction of a single clinical study to provide the appropriate PK data. Companies have therefore tended to perform clinical PK studies using the intravenous route (for drugs not otherwise intended for vascular administration) very reluctantly and only when regulatory authorities have demanded them.
Traditional Approaches to Intravenous Clinical Pharmacokinetic Studies The cost and resource challenges of intravenous administration to humans has traditionally limited the conduct of these types of study to situations in which a regulatory agency has asked for clinical absolute bioavailability data. A typical study involves a 2-way crossover design in which a single cohort of subjects is administered the drug extravascularly (typically oral) on 1 dosing occasion and intravenously on the other dosing occasion (separated by a suitable washout period). Diet, environment, and the consumption of other medications are controlled as far as
possible to minimize variables across dosings (ie temporal effects). The routes of administration may also be randomized between dosing occasions so that temporal effects can be minimized. Of course, it is not possible to entirely control these aspects, and the introduction of variables between dosing occasions that might affect the data cannot be ruled out entirely. Following dosing, the plasma–drug concentrationversus-time data are acquired for both dose routes, and the absolute bioavailability (F) is calculated by comparing the area under the drug concentration–time curve extrapolated to time-infinity (AUC0–1) for the extravascular (AUCex) and intravenous (AUCiv) doses, normalized to the doses administered (equation 1). F¼
AUC ex AUV iv
Doseiv Doseex
ðEquation 1Þ
Although equation 1 is commonly used in the determination of F, it should be remembered that it does not show the influence of CL. The 2 equations in respect to CL are shown in equations 2 and 3 below. CLiv ¼
Doseiv AUC iv
ðEquation 2Þ
CLex ¼
ðF Doseex Þ AUC ex
ðEquation 3Þ
Thus, F calculated from equation 1 assumes that CL is the same for both oral and intravenous administration. This is probably a reasonable assumption providing CL is not plasma drug–concentration dependent. The ideal crossover study design, therefore, is one in which the plasma concentrations (and hence AUCs) attained for the oral and intravenous doses are the same, or at least as close as possible. In practice, it is very difficult to design a study that will reliably achieve equivalent AUCs across the dose routes without already knowing F and V. There are cases, for example, in which apparent absolute bioavailability appears to be in excess of 100% because of membrane transporter–mediated nonlinear PK.11 Nonequivalent clearance is also thought to be responsible for an approximate 10% error in the determination of F for phenytoin.12 Error in the calculation of F arising from nonequivalent CL is rarely considered in absolute bioavailability studies, although the effect has been recognized for several decades. The crossover absolute bioavailability study design is limited to single-dose pharmacokinetics and does not take into account changes in V, CL, and F that might arise after repeat dosing. For chronically administered drugs in particular, it might be more appropriate to measure V, CL, and F under steady-state conditions. Calculation of V, CL, and F at steady state has been attempted by various
Lappin
13
Figure 1. The general principle behind the concomitant oral and intravenous isotopic tracer method to determine absolute bioavailability (F) and other PK parameters such as CL and V. The specific applications are explained in the text.
mathematical methods based on single-dose data. Vss, for example, is commonly calculated from the product of CL and mean residence time. Steady state conditions via the intravenous route can also be simulated for a limited time by a slow infusion. A method that experimentally measured these PK parameters under long-term steadystate conditions would be preferable, however.
Advent of Intravenous Clinical Pharmacokinetic Studies Involving Isotopic Tracers The issue of nonequivalent CL in the determination of F was recognized in the late 1960s to mid-1970s, which led to an alternative approach to the crossover design.13,14 Instead of administering the oral and intravenous doses on 2 separate dosing occasions, both doses were administered simultaneously, but the drug given intravenously was labeled with an isotopic tracer. In the first clinical study of this type by Strong et al in 1975,14 the absolute oral bioavailability (along with CL, V, and CLR) of N-acetylprocainamide (NAPA) was determined by administering a 500-mg single oral dose simultaneously with an intravenous dose of 250 mg 13C-NAPA (although the original study by Strong et al administered the oral and intravenous doses simultaneously, subsequent studies had a period of an hour or more between dose administrations, and so it is more correct to call the dosing concomitant;
this is explained further below). The plasma drug concentrations were independently determined using gas chromatography–mass spectometry for nonlabeled NAPA (from the oral dose) and 13C-NAPA (from the intravenous dose) by means of their different molecular masses. Using this method, both the oral and intravenous PK were determined from the drug concentration–time curves obtained from the analysis of a single set of plasma samples taken over time following the 1 dosing occasion. Because the overall systemic drug concentration was equal to the sum of that arising from the oral and intravenous doses, questions of nonequivalent kinetics that might otherwise occur between separate dosing occasions were virtually eliminated. In addition, this study design with its single dosing occasion eliminated temporal effects between dosings. The general principle of the isotopic tracer technique is illustrated in Figure 1. The study design employed by Strong et al had a particular shortcoming in that the intravenous administration of 250 mg made a significant contribution to the systemic drug concentration in addition to that delivered by the oral dose of 500 mg, which then had to be accounted for in the PK calculations (a multicompartmental model was used in this particular case). A better design would have been to administer a very small amount of drug by the intravenous route so that it did not significantly add to the plasma drug concentration derived from the oral dose. If the intravenous dose was kept sufficiently low, then it
14 would be reasonable to assume that the total plasma drug concentrations arose from the oral dose alone, thereby removing the need to deconvolute the kinetics. In the mid1970s, however, the sensitivity of analytical technology limited the degree to which the intravenous isotopically labeled dose could be lowered. The technique of Strong et al was adopted in a number of studies that followed utilizing a variety of isotopic tracers including 13C,15–19 15N,20,21 14C,22 18O,23 and 2 24 H. In addition, other approaches to estimating F were developed, including semisimultaneous dosing, in which an intravenous dose was given shortly after an oral dose (both nonlabeled) and the AUCs deconvoluted by mathematical modeling (known as Monte Carlo simulation).25,26 Alternative modeling methods attempted to estimate bioavailability from oral data alone.27,28 None of these methods were widely adopted, however, because of the inherent uncertainties of modeling compared with those based on an empirical approach.27,29 Moreover, there are uncertainties whether the regulatory authorities would accept such modeling studies as definitive data in a regulatory submission. The isotopically labeled intravenous dose approach first developed by Strong et al addressed potential issues of nonequivalent kinetics and removed questions of temporal effects between dose occasions, but the cost and effort surrounding regulatory approval for the intravenous administration remained, and so clinical absolute bioavailability studies still remained unpopular with the industry. However, the situation changed in 2006.
Advances in Isotopic Tracer Methods for Clinical Pharmacokinetics In 2006 the Australian Therapeutic Goods Administration Pharmaceutical Subcommittee announced that absolute bioavailability studies would be a requirement for all new chemical entities.4,7 As stated above, regulatory agencies might accept an argument that a particular study is unnecessary or impractical if a sound case can be made. The Australian agency’s announcement, however, made the exclusion of an absolute bioavailability study more difficult, and many companies were resolved to conducting an absolute bioavailability study as part of the registration package.7 It was also coincidently in 2006 that a publication appeared that reported the absolute bioavailability of 3 drugs using the isotopically labeled intravenous dose method (as depicted in Figure 1), in which the intravenous administration was only 100 mg each.30 For reasons that will become apparent, the isotope used in these investigations was 14C. This very low intravenous dose did not significantly contribute to systemic drug concentrations arising from the oral dose (oral doses ranged from 7.5 to 250 mg depending on the drug; see below) but demanded a
The Journal of Clinical Pharmacology / Vol 56 No 1 2016
very sensitive analytical method to determine the plasma 14 C-drug concentrations. The chosen method of 14C-drug analysis was accelerator mass spectrometry (AMS)31 that under the appropriate conditions can measure 14C in the low attomole range.32 The 3 drugs were erythromycin, midazolam, and ZK253 (a drug dropped from the Schering development pipeline because of poor oral bioavailability). In each case the drug (erythromycin, midazolam, or ZK253) was administered orally at a therapeutic dose level concomitantly with a 14C–isotopically labeled 100-mg intravenous dose. Plasma samples were analyzed with both liquid chromatography–tandem mass spectrometry (LC-MS/MS; tuned to the 12C ion) and by high-pressure liquid chromatography (HPLC) separation of parent drug followed by isotopic 14C AMS analysis. Using AMS enabled drug concentrations as low as 100 fg/mL plasma to be quantified. Around this time the regulatory agencies started to publish guidelines on the safety toxicology requirements necessary to allow the administration of very low doses of drug, termed a microdose (defined as 1% the pharmacologic dose or 100 mg, whichever is the lower). Microdoses are used in so-called phase 0 studies, principally to attain preliminary pharmacokinetic data in humans prior to phase 1. Because the toxicological risks associated with such low doses are inherently low, the regulatory guidelines allowed administration of a maximum of 100 mg on the basis of a much reduced safety toxicology package. Although a microdose and the intravenous administration of an isotopically labeled drug discussed in the current review are different concepts, the regulatory guidelines treated them very similarly. In fact, the guidelines called both types of administration a microdose, which has led to some considerable confusion and misinterpretations of data in the literature (examples are discussed in Lappin et al, 201333). Because of this, a new term has started to appear in the literature that refers to the intravenous dose of an isotopically labeled drug in an absolute bioavailability study administered as 1% of the pharmacologic dose or 100 mg, whichever is the lower, as a microtracer.34 The use of the term microtracer distinguishes the method discussed in this review from a microdose and hopefully will avoid future confusion in the terminology. This review uses the term microtracer henceforth. The most recent regulatory guideline, ICH M3, enables a microtracer to be administered intravenously concomitantly with a therapeutic oral dose without any additional safety data other than that required for the therapeutic oral dose plus corresponding systemic exposure data.35 Thus, by reducing the intravenous dose to a maximum of 100 mg, the isotopic tracer design not only essentially removed issues of nonequivalent kinetics and temporal effects but also eliminated the expense of conducting intravenous toxicology. In addition, the formulation of
Lappin
15
Table 1a. Summary of Microtracer Studies Published in the Peer-Reviewed Literature to Date Intended for Research or Method Development PK Parameters Determineda
Drug Erythromycin, Midazolam, ZK253 b-sitosterol R-142086 (Daiichi-Sankyo) Fexofenadine Clarithromycin Sumatriptan Propafenone
Isotope
i.v. Dose
CL, V, F
14
100 mg
CL, V, Vss, F, metabolic turnover CL, V, F
14
3-4 mg
14
100 mg
CL, V, Vss F CL, V, Vss, F
14
100 mg 100 mg
C
C C C C
14
Notes
Reference Lappin et al (2006)30
Microtracer over a dietary intake of b-sitosterol In dogs. Microtracer compared to cross-over
Duchateau et al (2012)43 Miyaji et al (2009)49 Lappin et al (2010)50 Lappin et al (2011)51
A microtracer is defined as an isotopically labeled intravenous administration of 100 mg or less given concomitantly with an extravascular dose. To the author’s knowledge, in all cases in which 14C drug and AMS were used, the levels of administered radioactivity were below those requiring regulatory approval, and therefore no dosimetry studies were necessary. Also to the author’s knowledge, intravenous doses were administered without additional intravenous toxicology data other than possibly an occasional local tolerability study for the intravenous injection site. a PK parameters relevant to the microtracer administration excluding t1/2 (and by implication Kel) and C0, which were obtained in all cases.
100 mg, or less of a drug for intravenous dosing is significantly less problematic than with higher concentrations. Indeed, some studies (indicated in Tables 1a and 1b) reduced the intravenous dose to much below 100 mg when the necessity arose. Formulation of such small intravenous doses, even with relatively insoluble drugs, can usually be achieved in vehicles such as physiological saline or glucose. Taking this approach also avoids risks of PK artifacts that might be caused by more esoteric dose
formulations.9 In cases of highly insoluble drugs, the dose can be lowered (eg,
