Therapeutic drug monitoring

1. What is therapeutic drug monitoring?
Therapeutic drug monitoring refers to the individualisation of dosage by maintaining plasma or blood drug concentrations within a target range (therapeutic range, therapeutic window). There are two major sources of variability between individual patients in drug response. These are variation in the relationship between:

• dose and plasma concentration (pharmacokinetic variability)
• drug concentration at the receptor and the response (pharmacodynamic variability) (Fig. 1).

By adjusting doses to maintain plasma drug concentrations within a target range, variability in the pharmacokinetic phase of drug action is greatly reduced. The major sources of pharmacokinetic variability are shown in Table 1.

2. For which drugs is monitoring helpful?
The characteristics of drugs which make them suitable for, or make them require, therapeutic drug monitoring are:

• marked pharmacokinetic variability
• concentration related therapeutic and adverse effects
• narrow therapeutic index
• defined therapeutic (target) concentration range
• desired therapeutic effect difficult to monitor

If the clinical effect can be readily measured (e.g. heart rate, blood pressure), it is obviously better to adjust the dose on the basis of response. Where this cannot be done, therapeutic drug monitoring is used in two major situations:

• drugs used prophylactically to maintain the absence of a condition such as seizures, cardiac arrhythmias, depressive or manic episodes, asthma relapses or organ rejection
• to avoid serious toxicity as with the aminoglycoside antibiotics which, unlike most antibiotics, have a narrow therapeutic range.

A list of drugs for which therapeutic drug monitoring is commonly used is shown in Table 2 with the target or therapeutic ranges. The ranges used are in most cases derived from observation of therapeutic and adverse effects in small groups of patients. Therefore, when applied to a wider population of patients, there will be individuals who achieve adequate effects at lower concentrations or experience adverse events within the 'therapeutic range'. There are two major important principles in using therapeutic ranges:

• Most drugs' responses are graded responses and are continuous through the concentration range (see Article 10 'Pharmacodynamics - the concentration- effect relationship' Aust Prescr 1995;18:102-4). Therapeutic responses do not magically 'switch on' at the lower limit of the therapeutic range nor do toxic responses suddenly appear at the upper limit.
• Individual patients will have individual therapeutic ranges - this is the residual pharmacodynamic variability in Fig. 1.

3. Sampling and drug analysis
Drug assay methods should have adequate sensitivity, be specific for the drug (or metabolite) to be measured and have appropriate accuracy and precision. Most high-volume drug assays are now carried out by automated immunoassay methods which have these characteristics. However, a number continue to require manual assay by methods such as high performance liquid chromatography (HPLC) and gas liquid chromatography (GLC) (e.g. amiodarone, perhexiline). National and international quality control programs are available for most commonly monitored drugs, and reputable laboratories are accredited by organisations such as the National Association of Testing Authorities.

Usually, plasma or serum is used for drug assays, depending on the equipment used. However, with cyclosporin there are large shifts of drug between red cells and plasma with storage and temperature change so whole blood is assayed. Some blood collecting tubes, especially those containing a gel to separate cells and plasma, may not be suitable for all drugs due to drug adsorption by the gel or other components in the tube.

Timing
The correct time of sampling is important. Drug concentrations vary over the dosing interval and with the duration of dosing in relation to achieving a steady state (see Article 11 'Designing dose regimens' Aust Prescr 1996;19:76-8). This is unlike most physiological parameters such as serum creatinine or serum sodium which change relatively slowly.

The least variable point in the dosing interval is the pre-dose or trough concentration. For drugs with short half-lives in relation to the dosing interval, samples should be collected pre-dose. For drugs with long half-lives such as phenytoin, phenobarbitone or amiodarone, samples collected at any point in the dosage interval can be satisfactory. For digoxin, any point after the distribution phase (after 6 hours post-dose - see Article 2 'Volume of distribution' Aust Prescr 1988;11:36-7) is acceptable. It should also be remembered that therapeutic ranges have often been established using trough concentrations.

Allowance will have to be made if samples are taken at other points in the dosage interval.

As outlined in Articles 2, 11 and 3 ('Half-life' Aust Prescr 1988;11:57-9), the approach to steady state is determined by the half-life and the use, or not, of a loading dose. It is usually best to wait and assay at steady state unless there are concerns about toxicity. This does not apply to drugs such as amiodarone and perhexiline with very long half-lives and which can cause severe toxicity - steady state may take months to be reached and dose adjustments need to be made along the way. With all drugs, if a sample is taken before steady state is reached, allowance needs to be made for this in interpreting the concentration.

4. What information is required for interpretation?
Drug concentrations need to be interpreted in the context of the individual patient without rigid adherence to a therapeutic range.

The basic information about the sample and the patient required for adequate interpretation of a drug concentration is shown in Table 3. Besides this, a good knowledge of the disposition of the drug is needed.

There are two important factors which can make interpretation of a result difficult in some cases. These are changes in protein binding and active metabolites.

Protein binding
Assays are done using plasma or blood and thus measure bound and unbound drug, whereas it is the unbound drug that interacts with the receptor to produce a response. If binding is changed by disease states, displacement by another drug or non-linearity in protein binding, the interpretation of total plasma or blood drug concentrations must be modified (Article 8 'Drug protein binding' Aust Prescr 1992;15:56-7). For example, the therapeutic range for phenytoin based on total drug concentration is 10-20 mg/L which corresponds to an unbound drug concentration of 1-2 mg/L (fraction unbound, f_u is normally 0.1). If f_u is increased to 0.2, as, for example, in renal disease, the target unbound concentration is still 1-2 mg/L, but the therapeutic range for total drug is 5-10 mg/L. Unless this is realised, inappropriate dose adjustments may be made, resulting in toxicity.

Sodium valproate and salicylate show non-linear binding in the therapeutic range making interpretation of total drug concentrations difficult (see Article 9 'Non-linear pharmacokinetics' Aust Prescr 1994;17:36-8).

Active metabolites
Metabolites which may not be measured can contribute to the therapeutic response. Examples include carbamazepine (carbamazepine-10,11-epoxide), and procainamide (N-acetylprocainamide). Theophylline in neonates (but not in adults) is converted to caffeine, so the therapeutic range for theophylline in neonatal apnoea is 6-12 mg/L (allowing for the contribution of caffeine), whereas it is 10-20 mg/L for obstructive airways disease in adults. The therapeutic ranges for imipramine and amitriptyline are based on the combined concentrations of parent drug and active metabolite (desipramine and nortriptyline respectively). Finally, primidone treatment is monitored by measuring the concentration of the active metabolite phenobarbitone, but primidone itself and another metabolite, phenylethylmalonamide, are also active.

5. Dose forecasting
Several methods have been developed to improve the prediction of individual dose requirements based on sparse data for individual patients. These are based either on calculation of clearance and volume of distribution from one or a few timed drug concentrations, or by a Bayesian feedback method. This latter method is based on differences between 'typical' population parameter values and those predicted for the individual patient from measured drug concentrations.

6. Is monitoring cost-effective?
Therapeutic drug monitoring is now so taken for granted that the difficulties in managing some drugs without it are forgotten. While the digoxin therapeutic range is somewhat 'loose', the advent of monitoring resulted in a far greater appreciation of the toxicity of digoxin and of the need for rational dosing. Similarly, theophylline, while now falling out of favour for other reasons, was rescued from oblivion when the advent of therapeutic drug monitoring in the 1970s allowed its use largely without the serious toxicity previously associated with it. The use of any of the drugs in Table 2 without monitoring would be difficult and often dangerous. Emphasis should be placed not so much on whether monitoring is necessary as on how to use it in the most cost-effective and clinically effective manner possible.