JOBNAME: tdm 27#1 2005 PAGE: 1 OUTPUT: Sat January 8 01:45:35 2005

lww/tdm/91753/04-009

REVIEW ARTICLE

Cost-Effectiveness of Therapeutic Drug Monitoring

A Systematic Review
D. J. Touw, PharmD, PhD,* C. Neef, PharmD, PhD,† A. H. Thomson, MSc, PhD,‡ and
A. A. Vinks, PharmD, PhD, FCP§ on behalf of the Cost-Effectiveness of Therapeutic
Drug Monitoring Committee of the International Association for Therapeutic
Drug Monitoring and Clinical Toxicology

principles have gained wide acceptance include the following:

Abstract: There are a number of effective but highly toxic drugs
that exhibit a narrow therapeutic index and marked interpatient
pharmacokinetic variability. Individualized therapy with such drugs
requires therapeutic drug monitoring (TDM) to obtain the desired
clinical effects safely. Cost-effectiveness analysis in health care is still
at an early stage of development, especially for TDM. A systematic
review was carried out to document studies that have addressed the
cost-effectiveness of TDM. The Cochrane database and Medline were
searched. References identified by this approach were then searched
manually for relevant articles. Very few studies have been performed
that document the cost-effectiveness of TDM, and TDM has been
demonstrated to be cost-effective only for aminoglycosides. For the
other classes of drugs that are monitored, the rationale for TDM has
been supported, but appropriate cost-effectiveness analyses have not
been performed. Because the use of many of these drugs without
TDM would increase the risk of under- or overdosing, emphasis
should not be placed solely on cost-effectiveness but rather on how
such interventions can be applied in the most cost-effective and
clinically useful manner.
Key Words: TDM, cost-effectiveness analysis, cost-benefit analysis
(Ther Drug Monit 2005;27:10–17)
(1) although imperfect, a better relationship often exists be-
tween the effect of a given drug and its concentration in the
blood than between the dose of the drug and the effect; (2)
a thorough understanding of pharmacokinetics, ie, the pro-
cesses of drug absorption, distribution, metabolism, and drug
excretion in individual patients and in patient populations is
available; and (3) the development of reliable and relatively
easy to use drug-monitoring assays. In addition, TDM can also
be useful in cases in which compliance is in question, where it
is not clear if the right drug is being taken, where dosage
adjustment is required as a result of drug–drug or drug–food
interactions, and where intoxication is suspected.
TDM is more than simply the analysis of a single drug
concentration in the blood of a patient and a report of this
number. It also comprises interpretation of the value measured
using the mathematical (pharmacokinetic) principles men-
tioned above, drawing the appropriate conclusions about the
result, and advising the physician who ordered the test how
to optimize treatment. It is important to apply a uniform
definition of TDM here, because different definitions have
previously been used in cost-effectiveness studies and reviews
of TDM. Consequently, comparisons have been made based
on different approaches, which may influence the results. The

T herapeutic drug monitoring (TDM) has been used to

individualize drug therapy since the early 1970s. In
addition, application of pharmacokinetic principles as part of
TDM has been used to assess drug behavior. The aim of TDM
is to optimize pharmacotherapy by maximizing therapeutic
efficacy, while minimizing adverse events, in those instances
where the blood concentration of the drug is a better predictor
of the desired effect(s) than the dose. The reasons why these
Received for publication January 23, 2004; accepted June 27, 2004.
From the *Apotheek Haagse Ziekenhuizen, 2504 AC Den Haag, Netherlands;
†Pharmacy Department, Medisch Spectrum Twente, Enschede, Nether-
lands; ‡Pharmacy Department, Western Infirmary, Glasgow, Scotland; and
§Pediatric Pharmacology Research Unit, Cincinnati Children’s Hospital
and Medical Center, Cincinnati, Ohio.
The authors have no vested interest in any of the products or methods
discussed.
Reprints: D. J. Touw, PharmD, PhD, Apotheek Haagse Ziekenhuizen, Postbus
43100, 2504 AC Den Haag, Netherlands (e-mail: d.touw@ahz.nl).
Copyright Ó 2005 by Lippincott Williams & Wilkins
International Association for Therapeutic Drug Monitoring
and Clinical Toxicology has adopted the following definition1:
n Therapeutic drug monitoring is defined as the measurement
made in the laboratory of a parameter that, with appropriate
interpretation, will directly influence prescribing procedures.
Commonly the measurement is in a biologic matrix of
a prescribed xenobiotic, but it may also be of an endogenous
compound prescribed as replacement therapy in an individual
who is physiologically or pathologically deficient in that
compound.
Besides a definition of TDM, clear definitions of cost-
effectiveness must be presented because most of the older
studies are financially driven. Useful definitions are:
Cost-benefit analysis: costs and benefits measured simulta-
neously in monetary units.
Cost-effectiveness analysis: costs and benefits measured
simultaneously in obtaining a specified objective and
in monetary units.

10 Ther Drug Monit  Volume 27, Number 1, February 2005

JOBNAME: tdm 27#1 2005 PAGE: 2 OUTPUT: Sat January 8 01:45:36 2005
lww/tdm/91753/04-009
Ther Drug Monit  Volume 27, Number 1, February 2005
Cost-utility analysis: benefits measured in terms of quality of
life, willingness to pay, and patient preference for one
intervention over another.
Cost-minimization analysis: costs analyzed and compared
when 2 or more interventions have been shown, or as-
sumed, to be equivalent in terms of outcomes.
In recent years, TDM has developed a much more
patient-oriented focus to include all the processes around drug
therapy (patient response, adverse events, dosing information,
blood draw times, pharmacokinetic behavior, drug analysis,
interpretation, and dose adjustment). The pharmaceutical
industry, in general, is not in favor of TDM because it is felt
that it increases drug-associated costs and raises barriers
against the success of their drug. However, in some cases,
well-conducted cost-effectiveness studies have shown the
opposite. TDM is not a goal in itself but aims to contribute to
the therapeutic drug management process, and a change from
the passive ‘‘monitoring’’ in the TDM acronym into (pro)-
active ‘‘management’’ has been advocated (C. Pippinger,
personal communication).
TDM and clinical pharmacokinetic services are usually
hospital-based services. These services can be provided by
either clinical pharmacists or clinical chemistry/pathology and
clinical pharmacology staff. TDM usually entails the reporting
of an analytic result in combination with an interpretation and
recommendation. These interpretations and recommendations
should be provided by adequately trained health care pro-
fessionals (eg, formal training in clinical pharmacology). In
times of increasing financial pressure for hospitals and their
budgetary constraints, objective measures of the cost-
effectiveness of a clinical pharmacokinetic service are required
to justify and maintain the staffing of this service.
There are several approaches to the evaluation of cost-
effectiveness. In 1966, Donabedian2 proposed the structure–
process–outcome-oriented method for the evaluation of the
quality of health care. Schumacher and Barr3 translated this
method to TDM. The structure component includes the total of
laboratory facilities personnel; the process component
comprises the process of proper sampling, interpretation,
and intervention based on the results; and the outcome
component comprises the clinical effect of the TDM
intervention, including speed of recovery, number of adverse
effects, morbidity, mortality, as well as cost-savings of TDM.
Schumacher and Barr identified the following categories for
TDM cost-effectiveness analysis3:
Patient oriented
What percentage of patients treated with TDM did not
experience adverse events?
Process oriented
What percentage of TDM blood samples was correctly drawn?
What percentage of TDM patients had serum concentrations
that were within a predefined ‘‘therapeutic or target
range’’?
It is obvious that patient-oriented studies are to be
preferred over process-oriented studies. In process-oriented
studies, surrogate endpoints are used because of the lack of
well-defined and measurable clinical parameters.
q 2005 Lippincott Williams & Wilkins
Cost-Effectiveness of Therapeutic Drug Monitoring
An important factor in carrying out a cost-effectiveness
analysis is the question of which costs and benefits are to be
included in the model. Are only the direct costs calculated
(human resources, materials, equipment, energy costs, main-
tenance, etc), or are indirect costs included (ie, costs associated
with loss of productivity of the patient, the costs borne by
relatives in supporting the patient, and healthcare costs asso-
ciated with a longer life). Benefits of TDM can be assessed in
several ways. Differences in mortality, morbidity, duration of
hospitalization, duration of drug therapy, number of drug
dosages, and number of serum concentrations ordered can be
costed and compared. Benefits can also be costed in natural
units (eg, costs per life year gained) or in qualys (quality-
adjusted life years). With this latter technique, each year that is
gained is also corrected for the quality of that extended life. It
is evident that well-conducted cost-effectiveness studies need
thorough pharmacoeconomic expert support.
Most of the published cost-effectiveness research in
TDM is process oriented, and many of these studies have been
summarized and discussed in an excellent review by Ensom
and co-workers.4 From this paper and other reviews it is clear
that the majority of TDM cost analysis studies have focused on
the aminoglycoside antibiotics.
Another important aspect of TDM is external pro-
ficiency testing. To date, most external proficiency-testing
programs have focused on testing the quality of the result in
relation to the assay methodology used. A logical next step is
also to evaluate the quality of the interpretation, or clinical
recommendation, together with the analytic aspects.5 Such
programs will contribute to ongoing quality improvements in
TDM and clinical toxicology.
This review aims to update the existing information on
the rationale and cost-effectiveness of TDM for different
classes of drugs by using the method of a systematic review.
METHODS
The medical literature was searched (MedLine) for the
indexed terms ‘‘cost-effectiveness’’ or ‘‘cost-benefit’’ or ‘‘phar-
macoeconomics’’ and ‘‘therapeutic drug monitoring’’ or ‘‘drug
analysis.’’ Studies found were selected for the categories
‘‘patient oriented’’ and ‘‘process oriented.’’ Reference lists of
review articles were searched by hand for studies that were
missed using the search terms.
Reviews and studies were rated according to the
following rating system [Anonymous. Canadian Task Force
Methodology (Canadian Task Force on Preventive Health Care
web site) available at http://ctfphc.org/methods.htm. Accessed
January 12, 2004]:
A1. Systematic review containing several studies of A2 level
and with consistent outcomes.
A2. Prospective randomized clinical trials of good quality.
B. Randomized clinical trials of moderate quality (eg, too
few patients) or other comparative trials (eg, not
randomized, cohort studies, case-control studies).
C. Noncomparative trials.
D. Experts’ opinions (eg, according to the authors).
A clinical trial was rated ‘‘good’’ if the study included
the following components: a clear hypothesis, the presence
11
JOBNAME: tdm 27#1 2005 PAGE: 3 OUTPUT: Sat January 8 01:45:36 2005
lww/tdm/91753/04-009
Touw et al
of blinding, analysis of confounders, statistical power and
sample size calculation, summary of patient characteristics,
specification of data analysis methods, and identification of
sources of bias such as the number of patients lost for
follow-up. In addition, a level of recommendation was as-
signed according to:
1. One systematic review (A1) or at least 2 independent trials
of level A2.
2. At least 2 independent studies of level B.
3. One study of level A2, B, or C.
4. Experts’ opinions, eg, according to the authors.
For each drug class, the review paragraph is summarized
by a conclusion and a recommendation.
Aminoglycosides
The success and continuing use of the aminoglycosides
can be attributed to various factors, including a rapid
concentration-dependent bactericidal effect, clinical efficacy,
synergism with b-lactam antibiotics, a low rate of true re-
sistance, and low costs. The major limitation to their clinical
use continues to be concern about the development of oto- and
nephrotoxicity. The nephrotoxic effects have been shown to be
related to active drug uptake and accumulation in the tubular
cortex, which is related to the dosing schedule. Therapeutic
drug monitoring (TDM) of the aminoglycosides is an
important tool that can reduce the incidence of toxicity while
maximizing efficacy parameters, such as the peak to minimal
inhibitory concentration (MIC) ratio. Several patient-oriented
studies have documented the cost-effectiveness of dose
individualization by means of TDM.6–11 As early as 1979,
Bootman et al6 demonstrated, in a retrospective cohort analysis
in burn patients with gram-negative sepsis, that dose individ-
ualization of gentamicin based on TDM increased the length
of stay in hospital but also significantly decreased mortality.
The extra costs of the TDM service were significantly less than
the overall savings made as a result of the reduced mortality.
These results have since been confirmed by other groups.7–9,11
More recently, research in Holland has demonstrated that
another approach to TDM, a more proactive approach in which
the hospital pharmacist is involved in deciding the initial dose
and dosing interval of an aminoglycoside for an individual
patient, can lead to a significant shortening of the hospital stay,
less nephrotoxicity, and a trend toward reduced mortality.10
This approach was compared with the passive form of TDM,
in which initial doses and sampling were decided by the
doctor. Based on hospital stay and additional medical costs,
the costs associated with the ‘‘active’’ TDM service were
significantly lower than the savings that were achieved by the
service. This is the only prospective European study to
evaluate the cost-effectiveness of TDM. An important
additional aspect of this study is that the cost-effectiveness
analysis was carried out by an independent pharmacoeco-
nomic research institute. The study has been archived in the
British National Health Service (NHS) Economic Evaluation
Database [National Health Service Economic Evaluation
Database (database online), York, Center for Reviews and
Dissemination, University of York, UK, http://www.york.ac.
uk/inst/crd/nhsdhp.htm. Updated January 5, 2004).
12
Ther Drug Monit  Volume 27, Number 1, February 2005
However, with the introduction of extended-interval
dosing for aminoglycosides, the role of TDM to assist dosing
has been the subject of considerable debate. There are no
clearly defined target concentrations for extended-interval
dosing, and the concept is difficult to apply in the conventional
sense.12 Trough concentrations should usually be below assay
detection limits and, if measurable, often imply overdosage.
The benefits of TDM services using population model-based
Bayesian pharmacokinetic approaches with extended-interval
dosing have not been formally studied. Based on the data
available, these strategies will have definitive advantages over
nomogram dosing in selected patient populations.
Panel 1: Conclusion and Recommendation on
Cost-Effectiveness With Respect to Aminoglycosides
Conclusion (level 1): TDM of aminoglycosides is cost-
effective [ref 6 (B), ref 7 (B), ref 8 (A2), ref 10 (A2)].
Conclusion (level 2): TDM of aminoglycosides leads to a
reduction of mortality [ref 6 (B), ref 10 (A2)].
Conclusion (level 2): TDM of aminoglycosides leads to a
reduction of side effects [ref 10 (A2), ref 11 (A2)].
Recommendation: It is recommended that aminoglycoside
therapy be guided by TDM.
Note: all cost-effectiveness studies reviewed here were per-
formed before the introduction of extended-interval
dosing.
Vancomycin
Vancomycin is considered to be less nephrotoxic than the
aminoglycosides. Nevertheless, a relationship seems to exist
between serum concentrations and both toxicity and efficacy.13
Two patient-oriented vancomycin TDM cost-effectiveness
studies have been conducted. The first study compared, pro-
spectively, the effect of TDM on the following outcome mea-
sures: renal function; cumulative dose; duration of vancomy-
cin therapy; duration of hospital stay; and consumption of
other antimicrobial drugs.14 On average, patients receiving
TDM had less nephrotoxicity, needed lower cumulative vanco-
mycin dosages, and had shorter hospital stays than patients
who did not receive TDM.
The second study was conducted by del Mar Fernández de
Gatta et al.15 They investigated the effect of vancomycin TDM
on clinical outcome and adverse effects. They also observed
a reduced incidence of mild and severe nephrotoxicity, and this
led to significant cost savings in patients receiving TDM.
Based on these 2 studies and a cost-price model for drug-
related nephrotoxicity, Darko et al calculated the savings that
could be realized by vancomycin TDM.13 Darko et al calcu-
lated the cost of treating nephrotoxicity as $11,233.13 The cost
of preventing 1 vancomycin-associated nephrotoxic episode
with TDM was $8363 for ICU patients, $5000 for oncology
patients, and $5564 for patients receiving concomitant nephro-
toxic drugs.13 Overall, TDM of vancomycin can be cost-effective,
with the biggest potential savings in ICU patients, patients
receiving other nephrotoxic medications, and oncology patients
because these patients have the greatest risk of developing nephro-
toxicity. Of more concern may be the potential for treatment
failure because of underdosing to avoid the risk of toxicity. There
q 2005 Lippincott Williams & Wilkins
JOBNAME: tdm 27#1 2005 PAGE: 4 OUTPUT: Sat January 8 01:45:37 2005
lww/tdm/91753/04-009
Ther Drug Monit  Volume 27, Number 1, February 2005
is preliminary evidence that treatment failures may result from
serum concentrations that are too low.16,17
Panel 2: Conclusion and Recommendation on
Cost-Effectiveness With Respect to Vancomycin
Conclusion (level 2): TDM is cost-effective in selected patient
populations: oncology patients, intensive care unit
patients, and patients treated with other nephrotoxic
drugs [ref 15 (A2), ref 13 (B)].
Conclusion (level 3): TDM of vancomycin results in reduced
nephrotoxicity [ref 15 (A2)].
Recommendation: It is recommended that vancomycin therapy
be guided by TDM in patient populations at risk, such as
ICU patients, oncology patients, and patients receiving
concomitant nephrotoxic medications.
Classic Antiepileptic Drugs
Since the 1960s, therapy with classic antiepileptic drugs
(phenobarbital, phenytoin, carbamazepine, primidone, valproic
acid) has been guided by TDM in Europe and in the United
States. However, well-conducted, patient-oriented studies on
the cost-effectiveness of TDM for antiepileptic drugs are
lacking. In addition, differences are felt for the need to use TDM
to optimize therapy. The need to monitor phenytoin, pheno-
barbital, and carbamazepine seems clear for pharmacokinetic
reasons, but there is less consensus for routine monitoring of
valproic acid and primidone. However, the published research
and consensus reports do not differentiate among the classic
antiepileptics. Because monitoring is so widely used in the
developed countries, clinicians may have acquired experience
in managing epilepsy with the aid of monitoring plasma con-
centrations, which may be a reason why the Italian TDM study18
found little difference in the outcomes of newly diagnosed
epilepsy, whether or not TDM was used. The main problem for
most TDM studies, and perhaps also for the Italian TDM study
on antiepileptic drugs, has been inappropriate study design. In
the Italian study, the investigators titrated the dose until serum
concentrations were in the therapeutic range and then com-
pared the results with those obtained from dosing based on
clinical endpoints.18 This is not what TDM is meant to be and
probably led to the conclusion that measurement of drug con-
centrations did not contribute to the outcome. To improve the
effects of TDM, initiatives have been developed to manage
TDM of antiepileptic drugs more rationally.19,20 Previously,
every patient on phenytoin received the same loading and
maintenance dose, and a dose was either skipped or added
according to drug concentrations. With their revised protocol,
they supplied initial dosage guidelines based on body weight
and provided an optimal sampling time for TDM. Further-
more, guidelines were provided as to what actions physicians
should take given the drug concentration result. Without being
a formal cost-effectiveness study, this new approach resulted in
a reduction in TDM requests and a higher efficacy of the
treatment (defined as better seizure control).20 When the
design and the results of the Italian TDM study group and of
Bates’s group are compared, it is striking that the outcome
depended on the expertise accompanying the TDM results.
Based on these results, Standards of Practice have been
formulated for the United States.21
q 2005 Lippincott Williams & Wilkins
Cost-Effectiveness of Therapeutic Drug Monitoring
Australian research has demonstrated better rates of
seizure control in patients monitored early in therapy (ie, in the
first 6 months of their epileptic disorder) but not if monitoring
began later in the course of treatment.22 This and older studies
have made it plausible that TDM can improve the effectiveness
of antiepileptic drugs.22
Recently, an Indian group has demonstrated that TDM
improves therapeutic outcome in adults with generalized tonic-
clonic epilepsy. Patients receiving TDM had better seizure
control, fewer adverse events, and a greater chance of remis-
sion. These patients also had a better social status (earning,
were married, had children) than patients who did not receive
TDM.23 The authors only compared the costs of the TDM
service based on the charges to the patients for the TDM
samples and did not take into account potential savings for the
community associated with the better social status of the
patients receiving TDM.
For the modern antiepileptic drugs (oxcarbazepine,
felbamate, gabapentin, lamotrigine, tiagabine, topiramate,
vigabatrin, zonisamide), there are no studies available that
have investigated the rationale and cost effectiveness of TDM.
Concentration–effect studies are scarce, and concentration–
effect relationships that have been described were established
in retrospective studies.24 For a number of new antiepileptic
drugs, pharmacokinetic parameters differ greatly between
individuals, and analysis of the serum concentration can
be helpful if a patient is difficult to control.24 In addition,
TDM can be useful to assess compliance and/or drug–drug
interactions.
Panel 3: Conclusion and Recommendation
on Cost-Effectiveness With Respect to
Antiepileptic Drugs
Conclusion (level 3): TDM of the classic antiepileptic drugs
can lead to better control of epileptic patients with fewer
side effects [ref 23 (B), ref 22 (D)].
Conclusion (level 3): TDM of the classic antiepileptic drugs
can be cost-effective [ref 23 (B)].
Recommendation: Therapy with the classic antiepileptic drugs
is preferably guided by TDM.
Conclusion (level 3): TDM of the modern antiepileptic drugs
can be useful in titrating patients whose epilepsy is
difficult to control and in cases of questionable
compliance and drug–drug interactions [ref 24 (D)].
Recommendation: Routine TDM of the newer antiepileptic
drugs appears not to be useful. TDM can be of help in
the titration and maintenance of patients who are
difficult to control.
Theophylline
Theophylline used to have an important place in the
treatment of asthma in the Western world. At present, it is used
principally in the treatment of COPD. There are no published
studies on the cost-effectiveness of TDM of theophylline.
However, there is a clear relationship between theophylline
concentrations and therapeutic and toxic effects.25,26 In addi-
tion, its pharmacokinetics show large interindividual variabil-
ity, and theophylline is very sensitive to drug–drug interactions
mediated by CYP1A2, its main metabolic pathway. Based on
13
JOBNAME: tdm 27#1 2005 PAGE: 5 OUTPUT: Sat January 8 01:45:37 2005
lww/tdm/91753/04-009
Touw et al
these considerations, TDM of theophylline is rational. A ret-
rospective study by Kearns et al has demonstrated that TDM
leads to higher and more appropriate dosing of theophylline in
children with asthma, without an increase in theophylline-
related toxicity.27
Panel 4: Conclusion and Recommendation on
Cost-Effectiveness With Respect to Theophylline
Conclusion (level 3): In view of the large interindividual
variability in theophylline pharmacokinetics, TDM is of
help in optimizing treatment [ref 27 (B), authors’
opinions (D)].
Conclusion (level 4): TDM of theophylline can be cost-
effective [ref 27 (D)].
Recommendation: TDM of theophylline therapy can be helpful.
Caffeine
Caffeine is used in neonates with apnea associated with
bradycardia and/or decreased oxygen saturation. Although
caffeine is frequently monitored, there are no studies that have
investigated the cost-effectiveness of routine caffeine moni-
toring in neonates. Recently, in an attempt to establish an
association between caffeine serum concentrations and clinical
efficacy, Keijer et al retrospectively studied the number of
apnea attacks in relation to caffeine TDM results.28 No
straightforward relationship between caffeine serum concen-
trations and the number of apneas could be demonstrated.
However, the authors commented that the retrospective nature
of the study had probably had a negative influence on the
registration of apneic periods.
Panel 5: Conclusion and Recommendation on
Cost-Effectiveness With Respect to Caffeine
Conclusion: Based on the present data, the rationale for caffeine
TDM cannot be established in neonates with apnea.
Recommendation: More studies are needed before a conclu-
sion can be drawn on the rationale for TDM in neonates
with apnea.
Digoxin
Digoxin is predominantly used in the treatment of atrial
fibrillation, but there are no published studies that document
the cost-effectiveness of TDM of digoxin in this condition.
Criteria for ordering a digoxin serum concentration have been
published recently.29 For use in atrial fibrillation, digoxin can
be titrated until sinus rhythm is achieved, and the serum level
(or even better, the peripheral compartment level30) serves as
the concentration at which the individual patient responds to
therapy. More recently, the use of digoxin in cardiac failure has
been under investigation. It was already known that digoxin is
effective in cardiac failure, but a reduction in mortality had not
been established.31 A post-hoc analysis of the DIG study32
demonstrated a positive effect of digoxin on mortality in men
with serum concentrations ranging from 0.5 to 0.8 mg/L.33
These findings can be questioned because this was a retro-
spective analysis, and the study was not designed to investigate
the effect of digoxin serum concentrations on mortality. A
strong point, however, is the large number of patients included
in this analysis. If this observation can be confirmed in
14
Ther Drug Monit  Volume 27, Number 1, February 2005
a prospective study, then TDM of digoxin is warranted because
of its very narrow target range, large interindividual variability
in pharmacokinetics, and potential for drug–drug interactions.
Panel 6: Conclusion and Recommendation on
Cost-Effectiveness With Respect to Digoxin
Conclusion (level 3): In cardiac failure, TDM of digoxin is
useful [ref 33 (B), authors’ opinions (D)].
Conclusion (level 4): In atrial fibrillation, TDM of digoxin
may be of use [authors’ opinions (D)].
Recommendation 1: In atrial fibrillation it can be clinically
useful to document the digoxin serum concentration
after titrating the patient with digoxin until the heart rate
is controlled.
Recommendation 2: When digoxin is used in cardiac failure, it
may be useful to use serum concentration measurements
to guide digoxin therapy.
Immunosuppressants
Immunosuppressants are used to prevent acute and
chronic organ rejection following transplantation. For cyclo-
sporine, tacrolimus, and sirolimus, blood concentration ranges
have been established that give the least chance of rejection
with acceptable side-effect profiles.34 Current immunosup-
pressants all exhibit large inter- and intrapatient variability in
pharmacokinetics, and in several concentration-controlled
trials it has been demonstrated that the blood concentration
is a better predictor of clinical efficacy than dose.35 From this
perspective, the immunosuppressants are ideal candidates for
TDM. Over the last decade, many consensus documents have
been published that address the need and methodology for
immunosuppressive drug monitoring; the most recent one
updating this information includes guidelines and recommen-
dations for cyclosporine, sirolimus, and tacrolimus.36 Never-
theless, there are no studies published that have formally
looked at the cost-effectiveness of TDM for immunosuppres-
sants. Because it has been clearly established that TDM
increases the chance of 1-year survival of a transplanted
kidney from 60% to 95%,35 a randomized study that inves-
tigates the cost-effectiveness of TDM versus no monitoring in
transplant patients would be ethically unacceptable. However,
studies that examine the best TDM strategies for different
transplant populations, for individual drugs, and for drug
combinations would be highly recommended. For example,
trough monitoring (C0), ‘‘peak’’ monitoring (C2), or abbre-
viated AUC monitoring may prove to be optimal for some
drugs in some circumstances.
In recent years, it has been demonstrated that dosing
based on the area under the curve (AUC) of cyclosporine
results in less rejection than dosing based on predose trough
concentrations. However, calculating the AUC requires that
serial blood samples be drawn at defined time points. This
approach, therefore, has practical limitations and is patient
‘‘unfriendly.’’ Recent studies in patients treated with cyclo-
sporine have shown the potential for utilizing sparse sampling
strategies in combination with a Bayesian approach. The
methodology achieved AUC estimates that were comparable to
those obtained using traditional sampling protocols, but fewer
samples were required. In addition, there was increased
q 2005 Lippincott Williams & Wilkins
JOBNAME: tdm 27#1 2005 PAGE: 6 OUTPUT: Sat January 8 01:45:38 2005
lww/tdm/91753/04-009
Ther Drug Monit  Volume 27, Number 1, February 2005
flexibility in the sampling times, which may have advantages
over more rigid approaches such as C2 monitoring (sampling
at a fixed 2-hour time point following dosing). Thus, the
approach is more patient friendly,37 and it offers the oppor-
tunity to integrate kinetics and dynamics when performing
cyclosporine TDM.
Panel 7: Conclusion and Recommendation
on Cost-Effectiveness with Respect to
Immunosuppressants
Conclusion (level 4): Because of a shortage of donor organs
and costs associated with rejection of a transplant,
immunosuppressants must be monitored. Because of
wide interindividual PK variability and risks for drug–
drug interactions, TDM of immunosuppressants is
required [authors’ opinions (D)].
Recommendation: Therapy with immunosuppressants must be
guided by TDM.
Psychiatric Drugs
There are no published studies on the cost-effectiveness
of TDM for psychiatric drugs. Only for a few psychiatric drugs
(lithium, imipramine, desipramine, nortriptyline, haloperidol,
and clozapine) has a relationship been described between
serum concentration and efficacy or toxicity.38–41 It must be
noted, however, that most investigators have looked for a direct
and linear correlation between the drug concentration in the
blood and clinical efficacy. Nonlinear analysis or PK-PD
modeling methods have seldom been used. Therefore, with
better analysis methods, more candidates for TDM are likely to
be identified. For many of the drugs listed, TDM will be
clinically useful because the wide interindividual variability in
pharmacokinetics that results from genetic polymorphisms of
drug-metabolizing (CYP-450) enzymes makes choosing the
appropriate dose difficult.
However, for most of the psychiatric drugs (apart from
those mentioned above), the so-called ‘‘therapeutic range’’ is
simply a reflection of the variability in the serum concentration
within the population when a standard dose is administered,
without considering whether the concentration is optimal or
effective. TDM is usually performed to verify patient com-
pliance, to identify differences in metabolism, eg, related to
genetic polymorphisms, or to investigate potential drug–drug
interactions.
Cost-effectiveness studies in psychiatry will be mostly
process oriented (eg, how fast is the result available, what
percentage of patients have a serum drug concentration that is
‘‘therapeutic’’). Patient-oriented studies could focus on the
question of what savings there might be for the community
when therapy is guided by TDM. Recently, it was found that
the chance of relapse and rehospitalization on clozapine
increased when the plasma concentration decreased more than
40% from baseline for more than 12% of the observation
period.42 The authors did not investigate any pharmacoeco-
nomic aspects of patient care in their study. Based on these
data, an intervention study could be designed that assesses the
value of maintaining the serum concentration within a target
range.
q 2005 Lippincott Williams & Wilkins
Cost-Effectiveness of Therapeutic Drug Monitoring
Panel 8: Conclusion and Recommendation
on Cost-Effectiveness with Respect to
Psychiatric Drugs
Conclusion (level 1): TDM is clinically useful for lithium,
nortryiptyline, desipramine, imipramine, haloperidol,
and clozapine [ref 38 (A1), ref 41 (C), ref 42 (C)].
Conclusion (level 4): For the other psychiatric drugs, TDM
can be of use in questions of compliance and drug–drug
interactions [authors’ opinions (D)].
Recommendation: Therapy with lithium, nortriptyline, de-
sipramine, imipramine, haloperidol, and clozapine
should be guided with TDM.
Protease Inhibitors
With the introduction of the protease inhibitors and non-
nucleoside reverse transcriptase inhibitors, the prognosis of
HIV-positive patients has improved dramatically. To date there
are no studies that investigate the cost-effectiveness of TDM
for HIV drugs. The protease inhibitors and the non-nucleoside
reverse transcriptase inhibitors show a wide inter-individual
variability in pharmacokinetics and are subject to drug–drug
interactions, and, for some, a correlation between serum
concentration and clinical efficacy has been established.43,44
Despite the use of these potent drugs, long-term therapy fails
in more than 50% of patients.45 The question as to whether
TDM can improve the prognosis for HIV patients can be
answered in part by the ATHENA study.46 Newly diagnosed
patients who were HIV drug naive were treated with nelfinavir,
and, in a randomized design, the effect of TDM on therapeutic
outcome was studied. Patients receiving TDM could be treated
for a longer period with nelfinavir before resistance developed
than patients treated without using TDM. Patients who had
TDM also experienced a lower incidence of therapy failure
with nelfinavir and had lower viral loads.46 Another study
showed clearly that failure to comply with therapy can be
identified by using TDM.47,48 Because poor compliance is the
main cause of failure with HIV therapy, TDM can contribute to
a greater chance of success in the long term. Despite these
positive findings, a pharmacoeconomic evaluation of TDM
has not yet been published.
Panel 9: Conclusion and Recommendations on
Cost-Effectiveness with Respect to HIV Drugs
Conclusion (level 3): TDM of nelfinavir is useful [ref 46 (A2)].
Conclusion (level 3): TDM of the other protease inhibitors and
nonnucleoside reverse transcriptase inhibitors can be
useful [ref 47 (C), ref 48 (C)].
Recommendation 1: Nelfinavir therapy must be guided by
TDM.
Recommendation 2: TDM can be useful to guide therapy for
the other protease inhibitors and nonnucleoside reverse
transcriptase inhibitors.
CONCLUSIONS
There is still a paucity of well-conducted studies inves-
tigating the added value and cost-effectiveness of TDM. The
importance and cost-effectiveness of TDM is well described
for aminoglycosides. For therapy with antiepileptic drugs,
15
JOBNAME: tdm 27#1 2005 PAGE: 7 OUTPUT: Sat January 8 01:45:39 2005
lww/tdm/91753/04-009
Touw et al
digoxin, psychiatric, and immunosuppressant drugs, TDM is
considered as standard of care despite the lack of formal cost-
effectiveness data. Without TDM, many of these drugs could
not be used as effectively, and there would be risks of either
underdosing or serious toxicity.
HIV drugs and chemotherapeutic agents (oncology)
have been identified as new areas for TDM.
The clinical efficacy of TDM is strongly correlated with
appropriate pharmacokinetic interpretation and is reflected in
individualized dosing recommendations being provided for
the clinicians who order the drug concentration analyses.
‘‘Numbers only’’ TDM services that provide test results
without appropriate interpretation and recommendations can
be misleading and do not use what we already know about
drug concentration–time profiles and behavior. Such TDM
services will predominantly generate costs without gaining
great clinical benefits. With the incorporation of sound rec-
ommendations to clinicians ordering drug assays, TDM has
already proven its added value in well-conducted studies.
Regarding the benefits that can be gained in terms of reduction
of mortality and morbidity and increase in medication safety,
one should question whether money should be spent on either
proving the cost-effectiveness of TDM or on increasing the
effectiveness of TDM.
REFERENCES
1. Watson I, Potter J, Yatscoff R, et al. Editorial. Therapeutic drug
monitoring. Ther Drug Monit. 1997;19:125.
2. Donabedian A. Evaluating the quality of medical care. Milbank Q.
1966;44:166–206.
3. Schumacher GE, Barr JT. Economic and outcome issues for therapeutic
drug monitoring in medicine. Ther Drug Monit. 1998;20:539–542.
4. Ensom MHH, Davis GA, Cropp CD, et al. Clinical pharmacokinetics in
the 21st century. Clin Pharmacokinet. 1998;34:265–279.
5. Vinks AATMM, Eerland JJ, Harteveld AR, et al. Analysis and dosing of
antibiotics in a renal patient: continuing education through EQUAS. Ther
Drug Monit. 1995;17:415.
6. Bootman JL, Wertheimer AI, Zaske D, et al. Individualizing gentamicin
dosage regimens in burn patients with gram-negative septicemia: a cost-
benefit analysis. J Pharm Sci. 1979;68:267–272.
7. Destache CJ, Meyer SK, Padomek MT, et al. Impact of a clinical
pharmacokinetic service on patients treated with aminoglycosides for
gram-negative infections. DICP. 1989;23:33–38.
8. Destache CJ, Meyer SK, Bittner MJ, et al. Impact of a clinical
pharmacokinetic service on patients treated with aminoglycosides: A
cost-benefit analysis. Ther Drug Monit. 1990;12:419–426.
9. Destache CJ, Meyer SK, Rowley KM. Does accepting pharmacokinetic
recommendations impact hospitalization? A cost-benefit analysis. Ther
Drug Monit. 1990;12:427–433.
10. Van Lent-Evers NAEM, Mathot RAA, Geus WP, et al. Impact of goal-
oriented and model-based clinical pharmacokinetic dosing of aminogly-
cosides on clinical outcome: A cost-effectiveness analysis. Ther Drug
Monit. 1999;21:63–73.
11. Streetman DS, Nafziger AN, Destache CJ, et al. Individualized pharma-
cokinetic monitoring results in less aminoglycoside-associated nephro-
toxicity and fewer associated costs. Pharmacotherapy. 2001;21:443–
451.
12. Vinks AA. Pharmacokinetic monitoring of aminoglycosides in the
extended dosing interval era. J Lab Med. 2002;26:43–48.
13. Darko W, Medicis JJ, Smith A, et al. Mississippi mud no more: Cost-
effectiveness of pharmacokinetic dosage adjustment of vancomycin to
prevent nephrotoxicity. Pharmacotherapy. 2003;23:643–650.
14. Welty TE, Copa AK. Impact of vancomycin therapeutic drug monitoring
on patient care. Ann Pharmacother. 1994;28:1335–1339.
16
Ther Drug Monit  Volume 27, Number 1, February 2005
15. Del Mar Fernández de Gatta M, Calvo V, Hernández JM, et al. Cost-
effectiveness analysis of serum vancomycin concentration monitoring in
patients with hematologic malignancies. Clin Pharmacol Ther. 1996;60:
332–340.
16. Zimmermann AE, Katona BG, Plaisance KI. Association of vancomycin
serum concentrations with outcomes in patients with gram-positive
bacteraemia. Pharmacotherapy. 1995;15:85–91.
17. Mulhern JG, Braden GL, O’Shea MH, et al. Trough serum vancomycin
levels predict the relapse of gram-positive peritonitis in peritoneal dialysis
patients. Am J Kidney Dis. 1995;25:611–615.
18. Jannuzzi G, Cian P, Fattore C, et al. A multicenter randomized controlled
trial on the clinical impact of therapeutic drug monitoring in patients with
newly diagnosed epilepsy. The Italian TDM Study Group in Epilepsy.
Epilepsia. 2000;41:222–230.
19. Schoenenberger RA, Tanasijevic MJ, Jha A, et al. Appropriateness of
antiepileptic drug level monitoring. JAMA. 1995;274:1622–1626.
20. Bates DW, Soldin SJ, Rainey PM, et al. Strategies for physician education
in therapeutic drug monitoring. Clin Chem. 1998;44:401–407.
21. Warner A, Privitera M, Bates D. Standards of laboratory practice:
antiepileptic drug monitoring. Clin Chem. 1998;44:1085–1095.
22. Eadie MJ. The role of therapeutic drug monitoring in improving the cost
effectiveness of anticonvulsant therapy. Clin Pharmacokinet. 1995;29:
29–35.
23. Rane CT, Dalvi SS, Gogtay NJ, et al. A pharmacoeconomic analysis of the
impact of therapeutic drug monitoring in adult patients with generalized
tonic-clonic epilepsy. Br J Clin Pharmacol. 2001;52:193–195.
24. Johannessen SI, Battino D, Berry DJ, et al. Therapeutic drug monitoring
of the newer antiepileptic drugs. Ther Drug Monit. 2003;25:347–363.
25. Burton ME, Vasko MR, Brater DC. Comparison of drug dosing methods.
Clin Pharmacokinet. 1985;10:1–37.
26. Holford N, Black P, Couch R, et al. Theophylline target concentration in
severe airways obstruction—10 or 20 mg/L? A randomised concentration-
controlled trial. Clin Pharmacokinet. 1993;25:495–505.
27. Kearns GL, Fischer TJ, Hunter RH. Use of serum theophylline deter-
minations during acute asthma therapy in children. Ann Allergy. 1982;48:
71–74.
28. Keijer WJ, Van Weissenbruch MM, Van Loenen AC, et al. Op weg naar
verfijning van de dosering. Farmacokinetiek van coffeine bij neonaten met
apneu. Pharm Weekbl. 2002;137:1021–1027.
29. Cañas F, Tanasijevic MJ, Máluf N, et al. Evaluating the appropriateness of
digoxin level monitoring. Arch Intern Med. 1999;159:363–368.
30. Kramer WG, Kolibash AJ, Lewis RP, et al. Pharmacokinetics of digoxin:
relationship between response intensity and predicted compartmental drug
levels in man. J Pharmacokinet Biopharm. 1979;7:47–61.
31. Riaz K, Forker AD. Digoxin use in congestive heart failure. Drugs.
1998;55:747–758.
32. The Digitalis Investigation Group. The effect of digoxin on mortality
and morbidity in patients with heart failure. N Engl J Med. 1997;336:
525–533.
33. Rathore SS, Curtis JP, Wang Y, et al. Association of serum digoxin
concentration and outcomes in patients with heart failure. JAMA. 2003;
289:871–878.
34. Johnston A, Holt DW. Immunosuppressant drugs—the role of therapeutic
drug monitoring. Br J Clin Pharmacol. 2001;52:61S–73S.
35. Shaw L, Kaplan B, Brayman L. Prospective investigations of concen-
tration–clinical response for immunosuppressive drugs provide the scien-
tific basis for therapeutic drug monitoring. Clin Chem. 1998;44:381–387.
36. Holt DW, Armstrong VW, Griesmacher A, et al. International Federation
of Clinical Chemistry/International Association of Therapeutic Drug
Monitoring and Clinical Toxicology Working Group on Immunosup-
pressive Drug Monitoring. Ther Drug Monit. 2002;24:59–67.
37. Cremers SCLM, Scholten EM, Schoemaker RC, et al. A compartmental
pharmacokinetic model of cyclosporine and its predictive performance
after Bayesian estimation in kidney and simultaneous pancreas-kidney
transplant recipients. Nephrol Dial Transplant. 2003;18(6):1201–1208.
38. Schou M, Juel-Nielsen N, Stromgren E, et al. The treatment of manic
psychosis by the administration of lithium salts. J Neurol Neurosurg
Psychiatry. 1954;17:250–260.
39. Amdisen A. Serum concentration and clinical supervision in monitoring
of lithium treatment. Ther Drug Monit. 1980;2:73–83.
40. Mitchell PB. Therapeutic drug monitoring of psychotropic medications.
Br J Clin Pharmacol. 2001;52(suppl):45–54.
q 2005 Lippincott Williams & Wilkins
JOBNAME: tdm 27#1 2005 PAGE: 8 OUTPUT: Sat January 8 01:45:39 2005
lww/tdm/91753/04-009
Ther Drug Monit  Volume 27, Number 1, February 2005
41. Ulrich S, Baumann B, Wolf R, et al. Therapeutic drug monitoring of
clozapine and relapse—a retrospective study of routine clinical data. Int J
Clin Pharmacol Ther. 2003;41:3–13.
42. Gaertner I, Gaertner HJ, Vonthein R, et al. Therapeutic drug monitoring of
clozapine in relapse prevention: A five-year prospective study. J Clin
Psychiatry. 2001;21:305–310.
43. Grub S, Delora P, Ludin E, et al. Pharmacokinetics and pharmacody-
namics of saquinavir in pediatric patients with human immunodeficiency
virus infection. Clin Pharmacol Ther. 2002;71:122–130.
44. Casado JL, Moreno S, Hertogs K, et al. NELSANE study. Plasma
drug levels, genotypic resistance, and virological response to a
nelfinavir plus saquinavir-containing regimen. AIDS. 2002;16:
47–52.
q 2005 Lippincott Williams & Wilkins
Cost-Effectiveness of Therapeutic Drug Monitoring
45. Back DJ, Khoo SH, Gibbons SE, et al. The role of therapeutic drug
monitoring in treatment of HIV infection. Br J Clin Pharmacol. 2001;52:
89S–96S.
46. Burger DM, Hugen PWH, Aernoutse RE, et al. Treatment failure of
nelfinavir-containing triple therapy can largely be explained by low
nelfinavir plasma concentrations. Ther Drug Monit. 2003;25:73–80.
47. Hugen PWH, Burger DM, Aarnoutse RE, et al. Therapeutic drug
monitoring of HIV-protease inhibitors to assess noncompliance. Ther
Drug Monit. 2002;24:579–587.
48. Hugen PWH, Langebeek N, Burger DM, et al. Assessment of adherence
to HIV protease inhibitors: Comparison and combinations of various
methods, including MEMS (electronic monitoring), patient and nurse
report, and therapeutic drug monitoring. J AIDS. 2002;30:324–334.
17