|
Analytical Issues in Therapeutic Drug Monitoring
By Ann Warner, PhD
Analyzing drug concentrations (TDM) has been part of the routine laboratory test menu only since the mid 1970s. Its development required two things: reliable analytical methods and a means of interpreting the results (pharmacokinetics).
With these two components in place TDM became one of the fastest growing areas in many laboratories in the 1980s. However, a recent study1 reported what has been documented before--that many drug concentration determinations are ordered improperly. Consequently, the data generated is clinically useless. In the study, the authors document that at their institution only 27 percent of anticonvulsant concentrations are ordered with a proper indication. This translated into an excess expenditure of $301,740 annually for useless TDM data.
Let's consider some of the specific issues which affect the quality of TDM results. Once laboratory scientists are aware of these, steps can be taken to increase both the analytical and clinical quality of the results which are generated.
Sample Timing
Samples are usually drawn at either peak (shortly after dose) or trough times (immediately prior to next dose). Peak concentrations are generally determined only for drugs dosed intravenously, due to the difficulties in determining when such a sample should be obtained for an orally dosed drug.
The drugs for which peak concentrations are most commonly determined are the aminoglycosides and vancomycin. This group of drugs is unique in that both peak and trough concentrations are measured when traditional dosing regimens are used (three times a day). However, when these three doses are combined into the so-called once a day dosing regimen, peak concentrations no longer need to be monitored. In this situation, the trough concentration is the most important one to evaluate in order to avoid toxicity.
By far the most common sampling time for the majority of drugs involved in TDM is immediately prior to the next dose, although some drugs which are dosed once in 24 hours may be sampled sometime in the middle of the dosing interval rather than waiting for the trough. What needs to be remembered is that there are several drugs which undergo a lengthy distribution phase after ingestion or infusion. Until this distribution is complete, the concentration of drug in the blood can appear to be at potentially toxic concentrations. This high concentration does not, however, correlate with the drug effect.
Drugs which have extended distribution phases include the following:
* Digoxin: minimum 6-10 hours, optimum 24 hours post dose
* Lithium: minimum 6 hours, optimum 12 hours post dose
* Cyclosporine: optimum immediately prior to next dose
Another factor which can lead to very elevated digoxin concentrations is the administration of Digibind®. A digoxin antidote, it binds to digoxin, inactivating it, and the complex is then excreted in the urine. Several immunoassays for digoxin also measure the Digibind-digoxin complex, producing very high concentration values which are not interpretable. As in the case of the improperly timed sample, the data is useless. Each laboratory can ascertain whether or not the assay they are using is affected by Digibind by checking the package insert or calling the company. The interference of Digibind can be removed through ultrafiltration.2
Practical Tips: Whenever a high digoxin concentration is encountered, particularly if it requires a dilution, it is a good policy to first call to determine if the sample has been collected too soon after a dose or Digibind has been used. The laboratory can conserve resources by not diluting and re-assaying such samples.
Some hospitals are addressing the timing issue by standardizing the times when a drug such as digoxin is dosed and samples are obtained for TDM.
Another factor to consider in determining when samples for TDM should be drawn is whether the drug concentration is affected by circadian rhythms. Such an effect would mean, for example, that a sample drawn 6 hours after the morning dose will have a different result from the sample drawn 6 hours after the evening dose. Particularly when a hospital patient is having his drug concentrations checked more frequently, consistent drawing times will help prevent unnecessary dosing adjustments. Drugs known to exhibit a circadian rhythm include carbamazepine, lithium, theophylline and cyclosporine.
Sample Site
It has been shown that even 6-12 hours after the immunosuppressive drug cyclosporine has last passed through an intravenous line, blood samples drawn from that line can be substantially elevated compared to samples from other sites.
Practical Tip: When a patient is receiving a drug intravenously, the best policy is to draw, if possible, from the alternate limb.
Types of Sample
The majority of TDM analyses are conducted using serum. The immunosuppressive drugs cyclosporine and tacrolimus (FK-506) require that the sample be whole blood because these drugs distribute between plasma and erythrocytes in a variable, temperature-dependent manner. Therefore, whole blood provides a better chance of obtaining comparable results from one sample to the next and between laboratories.
Serum is the sample of choice for most TDM. If plasma is used it is necessary to demonstrate the lack of interference by the anticoagulant on the chosen test procedure. With serum as the sample, a concern is whether or not tubes containing separator gels are appropriate for sampling.
Minimizing Effects of
Gel Containing Tubes
Gels contained in some separator tubes are known to adsorb some drugs, with the result that the measured drug concentration is falsely lowered. Such an effect is of concern when it causes a 10 percent or greater decrease in the resulting concentration.
Two steps can be taken to minimize the effects of separator gels: fill tubes completely, and process ssamples promptly.3,4
Practical Tip: Draw all samples for TDM which require serum in plain red top tubes.
Unbound (Free) Drug Monitoring
Drugs bind variably to plasma proteins. Some drugs are highly bound, others little or not at all. The portion of the drug which is bound to protein is inactive. The active portion of the drug is that portion not bound to protein which is free to cross membranes and interact with drug receptors to cause the drug effect.
When a drug is bound to protein there is a relationship between the free and bound drug amounts which depends in turn on the amounts of protein and drug present. As a result, changes in the amount of protein can alter the amount of free drug.
For highly bound drugs such as phenytoin and valproic acid, a relatively small decrease in the amount of protein available for binding may have a correspondingly large effect on the amount of free drug, as will be illustrated shortly. When TDM is performed, the total drug is what is measured. There are some cases in which the total drug concentration may be misleading and a free drug concentration is needed.
There are two main criteria for determining when measuring unbound drug concentration may be useful:
* The drug is bound to plasma proteins in excess of 85 percent to 90 percent (if a drug is only bound 50 percent to 70 percent, changes in binding will translate into only small changes in free drug concentration)
* The extent of binding of the drug can be variable as a result of changes in protein and/or drug concentration
Proteins which are responsible for the majority of drug binding in plasma are albumin, which binds mainly acidic drugs, and a1-acid glycoprotein, which binds basic drugs. In addition, pregnant women, the elderly and neonates have decreased albumin and/or altered protein binding.
For a drug such as phenytoin which is 90 percent bound to albumin, the amount of free drug will be 10 percent of the total concentration. If a 10 percent decrease in binding were to occur, due, for example, to a decreased amount of albumin, the following concentration changes would occur if the patient's dose is kept the same:
Measured
% Bound Total Drug (CT) Free Drug (CF)
90 20 µg/mL 2.0 µg/mL
80 10 µg/mL 2.0 µg/mL
In this example, as binding decreases from 90 percent to 80 percent, the free drug fraction, referred to as a , increases to 20 percent. The actual free drug concentration remains the same (as does the patient's response to the drug) but the total concentration decreases. If, as a result of the low total drug concentration, the clinician decides to adjust the patient's dose to produce a total drug concentration of 20 µg/mL, the following will occur:
Measured
% Bound Total Drug (CT) Free Drug (CF)
90 20 µg/mL 2.0 µg/mL
80 20 µg/mL 4.0 µg/mL
Because of the change in binding, the amount of free drug concentration is doubled and may cause toxic effects even though the total drug concentration is within the therapeutic range.
Example: A patient who is receiving phenytoin, a drug which is highly bound (90 percent, a = 0.1) to albumin is found on admission to the hospital to have an albumin concentration of 2.5 g/dL. When this patient's phenytoin concentration is checked it is found to be 6.1 µg/mL (therapeutic range = 10-20 µg/mL). Should this patient's dose be adjusted to a higher dose?
Answer: In a patient with a normal albumin concentration, a phenytoin concentration of 6.1 is likely to be sub-therapeutic. However, this patient has a markedly decreased albumin concentration which will affect the free drug concentration. An equation has been developed which allows for the measured total drug concentration at the lowered albumin concentration to be corrected to the equivalent total concentration if the patient's albumin were within the reference interval. Using this equation, the patient's adjusted total concentration can be calculated as follows:
Equation5:
CTN = (CTA) / [(1-a)(PA/PN) + a]
CTN = Total drug concentration with normal binding present
CTA = Total drug concentration with abnormal binding
PA = Patient's albumin concentration
PN = Normal albumin concentration,
e.g. 4.4 g/dL
a = free fraction of drug (if drug is bound 90%, a=10% or 0.1)
Patient Data:
CTN = 6.1 µg/mL / (1 - 0.1)(2.5 g/dL/4.4
g/dL) + 0.1
CTN = 6.1 µg/mL / (0.9)(0.568) + 0.1
CTN = 6.1 µg/mL / 0.61 = 10 µg/mL
Since the calculation indicates that the patient's current phenytoin concentration of 6.1 µg/mL is equivalent to a total concentration of 10 µg/mL if the albumin concentration were normal, the dose probably does not need to be adjusted.
Another way to evaluate patients with low albumin concentrations is to measure the free drug concentration directly, using ultrafiltration.
Key Concept: Routine drug monitoring involves the measurement of total (bound plus unbound) drug concentrations.
Practical Tip: An elevated free fraction of drug is frequently seen when albumin is <2.5 g/dL but is rare if albumin concentration, and hepatic and renal function are all normal.
Analytical Procedures for
Free Drug Determinations
Free drug concentrations can be measured directly following an ultrafiltration procedure which produces an ultrafiltrate of the serum containing only the free drug. This requires:
* An appropriate filter (e.g. Centrifree Filters®, Amicon Inc., Beverly, MA) to remove protein bound drug;
* Use of a fixed angle centrifuge to improve the efficiency of the ultrafiltration device;
* Constant temperature during centrifugation (temperature changes affect the amount of free phenytoin)6;
* Adjustment of the assay sensitivity in order to measure the lower concentrations produced.
Additional points to consider are that samples should be ultrafiltered prior to freezing, and serum is the sample of choice, since some anticoagulants can alter free fractions of some drugs.
Key Concept: Phenytoin is the drug most frequently monitored as a free drug.
Metabolites
Two major questions need to be asked about metabolites:
1.) Is the metabolite active, with either the same or a different pharmacologic action as the parent?
2.) Can it be measured?
A number of drugs, including procainamide and primidone, form active metabolites which can be analyzed and are, in fact, routinely measured along with the parent drug. There are also examples of drugs which form toxic, active and/or inactive metabolites which may affect the measurement of the parent, but which cannot be readily measured as individual entities (see Table 3).
It also has been demonstrated that patients with severe liver and kidney function problems may produce unusual metabolites which can have a variable effect on the concentration measurement, depending on the method used to quantitate the drug.
Summary
With the pressures being placed on health care to be more cost-effective, we are being forced to focus more and more on the issue of proper utilization of laboratory resources. In TDM this means we should be attempting to assure that the high analytical quality of our results is matched by their clinical utility. The concentration values, which we can produce with marvelous precision and accuracy in a relatively short time, are only useful clinically if the specimen was collected appropriately.
The result we produce will often be entered into a mathematical formula to verify appropriate dosing has occurred or to determine changes needed. If any of the information used in the formula is incorrect, the result may well be that the patient will be either under- or overdosed. The laboratory scientist responsible for TDM can and should take steps to ensure both analytical and clinical quality.
* About the author: Dr. Warner is a professor in the Department of Pathology and Laboratory Medicine at the University of Cincinnati (OH) Medical Center, and interim director of Toxicology at University of Cincinnati Hospital.
References
1. Schoenenberger, R.A., Tanasijevic, M.J., Jha, A., Bates, D.W. Appropriateness of antiepileptic drug level monitoring. JAMA. 1995;274:1622-6.
2. Ujhelyi, M.R., Cummings, D.M., Green, P., et al. Effect of digoxin Fab antibodies on five digoxin immunoassays. Ther Drug Monit, 1990;12:288-292.
3. Dasgupta A, et al. Absorption of therapeutic drugs by barrier gels in serum separator blood collection tubes. Am J Clin Path. 1994;101(4): 456-61.
4. Landt, M., Smith, C.H., Hortin, G.L. Evaluation of evacuated blood-collection tubes: effects of three types of polymeric separators on therapeutic drug-monitoring specimens. Clin Chem. 1993;39:1712-1717.
5. Winter, M.E. Basic Clinical Pharmacokinetics, 3rd Edition, 1994. Vancouver, Applied Therapeutics Inc.
6. Ratnaraj, N., Goldberg, V.D., Hjelm, M. Temperature effects on the estimation of free levels of phenytoin, carbamazepine and phenobarbitone. Ther Drug Monit 1990;12:465-72.
|