Testing drug levels

When someone does not respond to an antiretroviral regimen, any or all of the following could play a part:

• Poor treatment adherence resulting in low drug trough levels.

• Drug interactions that increase or decrease absorption.

• Genetic influences on an individual’s drug metabolism.

• Other varying pharmacokinetics among individuals (e.g. age, weight, height) that can result in suboptimal drug levels or toxicity/severe side-effects from drug levels that are too high.

Therapeutic drug monitoring (TDM) is the term used to describe the measurement of plasma drug concentrations. It is done to guide the dosing regimen for a patient and can help determine why someone is not doing well on therapy.

For several antiretroviral agents, researchers have demonstrated a relationship between drug levels and efficacy and/or toxicity. Several characteristics make certain drugs more suitable than others for therapeutic drug monitoring. The ideal drug for TDM should have:

• A clearly defined relationship between plasma concentration and pharmacologic effect and/or toxicity.
• Significant inter- and intra-individual variability in the drug plasma concentration at a fixed dose.
• A narrow therapeutic window (therapeutic index), that is, the level at which the drug is effective is very close to the level at which it is toxic.
• A high toxicity profile (dose-limiting toxicity).

A number of studies suggest that the particular characteristics of protease inhibitor (PI) and non-nucleoside reverse transcriptase inhibitor (NNRTI) classes of antiretroviral drugs lend themselves favourably to TDM. The nucleoside reverse transcriptase inhibitors (NRTIs) are less amenable to therapeutic drug monitoring as blood levels do not seem to correspond to active drug concentrations found in cells.

Antiretroviral drugs (ARVs) are processed into inactive products relatively quickly, hence the need for frequent dosing. Their metabolism is dependent on the liver, involving a 'pathway' called the cytochrome P450 system (CYP450), which is responsible for processing many other drugs and nutrients. Interactions between these different substances can affect the speed at which they are metabolised, causing blood levels to rise or fall.

The activity of CYP450 is itself variable. Some people are rapid CYP450 metabolisers and others are slow. This produces variation between individuals in protease inhibitor (PI) levels in the blood. Genetic tests are being developed in an attempt to identify rapid and slow metabolisers.

Poly-glycoprotein (P-gp) is a protein found on the surface of cells. P-gp acts as a transmembrane pump, removing drugs from the cell membrane and cytoplasm. Research has shown that P-gp inhibitors, such as the drug cyclosporin, can inhibit the transport of PIs, so the action of P-gp would seem to affect the amount of drugs found in cells.

There is currently much interest in the field of genetic differences (polymorphisms) in enzymes and transporters and how this affects serum concentrations of antiretroviral drugs. The hope is that genetic screening may be utilised to establish how individual variation may affect blood concentrations of ARVs. The concept is similar to the practice of genetic screening to identify patients at particular risk of abacavir hypersensitivity reactions.

CYP2D6 is the best characterised of the CYP isozymes, with more than 75 variants currently identified. While it accounts for only 2 to 5% of all CYP isozymes, CYP2D6 metabolises approximately 25% of all clinically used medications. Absent or reduced CYP2D6 activity results in drug accumulation because of reduced drug clearance efficiency.

While considerable information is being gathered in this field, more work is needed to develop practical applications for the role of genetics in predicting drug response (pharmacogenetics).

In order to decide how and when therapeutic drug monitoring might be useful in a practical sense, it is important to understand the benefits and limitations of this technique.

At present, British treatment guidelines and the US guidelines recommend therapeutic drug monitoring (TDM) when drug concentrations are difficult to predict. These settings include:

• In the management of suspected drug interactions.
• Pregnancy.
• Paediatrics.
• Malabsorption.
• Extremes of body mass index.
• Salvage therapy when TDM and resistance test results can be integrated.
• Patients with renal or hepatic impairment.
• Transplant patients.
• Cases of toxicity or concentration-dependent toxicities.
• Suspected non-adherence.
• The use of alternative dosing regimens whose safety and efficacy have not been established.
• Lack of expected virologic response in a treatment-naive person.

In order to measure the amount of drug in the blood a small quantity (at least 1ml) of plasma is needed. Two samples will often be taken in order to establish the highest and lowest concentration of drug in the body over a dosage interval. A peak sample is usually taken two hours after the last dose and a trough sample is collected at the end of the dosing interval. For example, in a twice-daily regimen, the trough sample would be done at hour 12 after dosing. If it is not practical for two samples to be collected, a single trough sample will usually suffice.  

If it is difficult to obtain a trough sample, an option is to use a randomly collected sample, which will allow a calculation of predicted trough and peak concentrations.

In Europe, a quality control system for TDM was established following studies that showed considerable variability in the quality of drug-level testing. Nijmegen University Medical Centre in the Netherlands organised the first international comparison study to determine how much variation exists in drug level measurements from different laboratories.

The researchers sent plasma samples from HIV-negative volunteers, spiked with three different, standardised doses of indinavir, nelfinavir, ritonavir, and saquinavir, to ten laboratories worldwide. Measurements were considered correct if they fell within 20% of the concentration of drug that should have been present, based on the original weight of drug dissolved in the plasma sample. This method was chosen to minimise interpersonal variation in drug metabolism.

Indinavir was the drug with the most consistent detection rate; 86% of measurements were correct. In contrast, ritonavir was poorly measured; only 42% of measurements were correct. For saquinavir and nelfinavir, two-thirds (65%, 67%) of measurements were correct.

Only one laboratory measured all three doses accurately. Overall, the ability to detect low concentrations correctly was worse than for other concentrations.⁹ Despite the significant levels of inaccuracy, the Dutch study showed a marked improvement in quality as compared to a prior study of nelfinavir drug-level monitoring in the US. That study looked at results from five laboratories and found a fivefold variation between labs.

Researchers and pharmaceutical companies conduct various studies in order to determine the pharmacokinetic parameters of drugs . Two such parameters are C_min and IC. C_min is the minimum concentration of a drug measured between one dose and the next. Inhibitory concentration (IC) values refer to the concentration required to inhibit wild-type virus (virus without resistance mutations).

Different ICs can be measured. IC₅₀, IC₉₀, and IC₉₅ refer to the concentration of a drug needed to inhibit viral replication by 50%, 90%, or 95% respectively. The IC values rise as virus becomes more resistant, reflecting the need for more drug to inhibit resistant virus. The IC value for one virus may be very different from the IC value for another. 

The ratio between the C_min and the IC₅₀ gives a value called the inhibitory quotient (IQ). This value is a way of illustrating numerically the comfort zone that exists when a particular drug is used. For example, if the trough level (C_min) of a drug is 50 and the IC₅₀ of a drug is 10, the ratio will be five. In other words, the trough level is five times higher than the minimum concentration needed to inhibit 50% of virus replication.

There are two types of IQs that can be calculated for resistant virus. Data collected from phenotypic resistance tests are phenotypic inhibitory quotients (PIQs) and data derived from genotypic analysis refer to the genotypic inhibitory quotient (GIQ). (See Resistance testing for more information on this topic.) This distinction between genotype and phenotype may be important. The phenotypic IC₅₀ should be adjusted for the way in which the analysed drug is transported in the body. Accordingly, an adjustment is needed to account for protein binding. At present, allowances made for protein binding differ from assay to assay.

Genotypic resistance testing has the advantage of being less expensive, but it also has limitations in terms of its application to determining IQ. For example, the model used to determine GIQ assumes an equal importance for each resistance mutation included. In reality, there may be differences. To date, there is no gold standard against which to weigh the importance of different mutations, so GIQ might vary considerably according to the various mutations that are present.

There have not been any prospective studies conducted for the validation of cut-off values for IQ-based TDM values or to validate the strategy of IQ-based TDM. At present, IQ use is considered experimental and its application is primarily limited to the clinical trial setting.