Using a blood test to monitor concentrations of antiretroviral drugs may be a clinical tool whose time has come. With the growing awareness of the toxicities and low tolerability of currently available anti-HIV drugs, a method for optimizing their use has become increasingly attractive. Clinicians already routinely use drug level monitoring, also known as therapeutic drug monitoring (TDM), in treating diseases such as epilepsy and asthma. But is TDM appropriate for HIV clinical care, or is its potential application too limited to warrant much enthusiasm? There is reason to believe, for instance, that TDM will not be used widely with all three classes of antiretroviral drugs. Several studies have nevertheless indicated a possible benefit to using TDM in HIV care, and a few small trials evaluating its efficacy in the clinic are underway. Until conclusive trial data become available, the future role of TDM in the treatment of HIV disease will remain unclear.

The Therapeutic Window

TDM is a type of blood test mainly used to monitor the concentration of a drug in the plasma (blood) to ensure that appropriate drug levels are maintained. In the case of antiretroviral agents, a certain amount of drug must always be present in the plasma to adequately inhibit HIV replication and keep the viral load at the lowest possible level. The range of drug concentrations that will safely achieve this result is known as the "therapeutic window." If the drug concentration falls below this range, the viral load may rise, increasing the possibility that mutant (genetically altered) forms of the virus will be generated, and that the drug -- or even an entire drug class -- will lose its efficacy. If the drug level rises above the recommended therapeutic window, toxic side effects and long-term complications might result.

Unfortunately, the therapeutic window for several anti-HIV drugs is narrow. Even the minimum levels of many anti-HIV drugs needed to suppress viral replication below the limit of detection can cause unpleasant adverse effects, such as peripheral neuropathy (burning or tingling sensations in the hands and feet) or diarrhea (see "Adverse Effects Associated with Antiretroviral Therapy" in BETA, Spring 2000). More importantly, approximately 50% of people taking the recommended doses of highly active antiretroviral therapy, or HAART, are not able to maintain an undetectable viral load, indicating insufficient drug exposure, drug resistance, or both. (Insufficient drug exposure may be caused by problems with adherence, or following a prescribed treatment regimen, which is discussed below.)

A limited range of therapeutically sound dosages is one of the reasons why the currently approved antiretroviral drugs are often difficult to take and of uncertain utility over a long period of time. The fact that the minimum drug concentration needed to suppress viremia (virus in the blood) is uncomfortably close to the maximum tolerated dose also suggests a critical need to optimize, or individualize, drug doses. Before discussing the complexities of drug individualization, it is worth examining how standard doses are derived.

Not Your Average Dose

The range of optimal drug levels is normally determined for an individual drug during the clinical trial process. But because antiretroviral drugs tend to have a slender therapeutic window, researchers have not been able to assign reliable target concentrations that are therapeutically appropriate to each person who needs these medications. Researchers must settle on a "standard dose" for a particular antiretroviral drug that reflects merely an average positive response to the drug that was seen among study subjects. Responses vary widely -- even in clinical trials that have led to Food and Drug Administration (FDA) approval, some trial participants were never able to achieve optimal plasma concentrations. Additionally, most study subjects to date have been antiretroviral-naive (not previously treated) adult White males, which further undermines the universal relevance of standard drug doses since several important factors such as age, race, gender, and treatment history can affect an indiv-idual’s response to a drug. In the past few years, however, clinical trials have included an increasing percentage of women, Latinos, and African-Americans.

The principal clinical role of TDM would be to derive optimal drug doses corresponding to the profile of the individual being monitored, not to an average response detected during clinical trials. TDM thus seems to be a logical step in the process from developing a drug to administering it in the individual. It is commonly used for drugs such as tobramycin (Tobrex or TOBI, a broad-spectrum antibiotic) and digoxin (Lanoxin, used in the management of congestive heart failure and some dysrhythmias, or abnormal heart rhythms). Genotypic drug resistance tests, which measure viral mutations, and phenotypic drug resistance assays, which measure resistance to individual drugs, could be used in tandem with drug monitoring to help physicians determine for individual patients which drugs and optimal doses would most effectively suppress HIV replication and minimize drug toxicity. (For more information about resistance testing, see "Genotypic and Phenotypic Resistance Testing" in BETA, Summer 1999.) The best timing or sequence of such a battery of tests is not yet known.

In several European countries, including France and the Netherlands, drug monitoring is already offered in clinics in a research setting, no doubt as a result of trials showing a correlation between individualized doses of protease inhibitors, or PIs (see next section) and improved clinical outcomes. In a 1998 article in AIDS, for example, Richard M.W. Hoetelmans, PharmD, PhD, of Slotervaart Hospital in Amsterdam and colleagues reported that HIV clearance rates were found to correspond with plasma concentration ratios of nelfinavir (Viracept) and saquinavir (Fortovase). Another Dutch study appearing the same year in Antiviral Therapy ascribed treatment breakthrough (defined as a viral load above 200 copies/mL after 24 weeks of treatment) to low plasma levels of indinavir (Crixivan) in persons taking triple-drug therapy. However, other trials have not shown a consistent correlation, suggesting a need for further investigation.

TDM and PIs

Even if TDM turns out to be a viable treatment strategy, individualizing drug doses for multidrug regimens will not be easy. As mentioned above, TDM probably cannot be used for all anti-retroviral drugs. Plasma concentrations of nucleoside reverse transcriptase inhibitors (NRTIs) are often not a marker of antiviral effect, primarily because these drugs must be converted within living cells to triphosphate anabolites (active forms) before they exert their antiviral effects. Triphosphate anabolites can be measured in peripheral blood mononuclear cells, or PBMCs (a type of white blood cell), but NRTI triphosphate research assays are nonstandardized, prohibitively expensive, and likely to remain too complex for widespread use. Additionally, the prolonged half-life (time needed for half of an original amount to be eliminated or metabolized) and relatively stable intra-dosing plasma concentrations achieved by non-nucleoside reverse transcriptase inhibitors (NNRTIs) indicate that this class of drugs may play a very limited role in subthera-peutic drug exposure (i.e., drug exposure that is too low to be useful).

There are, however, compelling reasons for using TDM to individualize regimens containing PIs, as PI plasma concentrations tend to correlate with antiviral effect and the half-lives of PI drugs tend to be short. Nevertheless, managing PIs with drug level monitoring presents several challenges.

Unresolved Pharmacokinetic Issues

Many factors influence the pharmacokinetics (absorption, distribution, metabolism, and excretion) of medications in the body. The genetic disposition of the individual plays a significant part. As mentioned above, different populations (e.g., African-Americans, Whites, children, and women) may exhibit different pharmacokinetic characteristics, but even people within the same population group may vary greatly in this regard. Weight and age of the individual, for example, can be pertinent factors in evaluating drug levels. In addition, drug absorption can be influenced by anti-HIV treatment history and stage of HIV disease -- for instance, people with more advanced disease may need higher drug doses to offset problems of drug absorption in the stomach and intestines. Liver dysfunction, due to chronic hepatitis B (HBV), hepatitis C (HCV) infection, and/or chronic alcohol abuse, can have a discernable effect on PI concentrations. Renal (kidney) function and clearance (drug elimination) should also be considered. Whatever the cause, PI levels between individuals may vary by more than 10-fold.

The individual who takes anti-retroviral drugs may also show remarkable variability over time, and even during the course of a single day. This may be explained by genetic, hormonal, as well as lifestyle factors. Diet -- both what is eaten and the timing of meals -- greatly influences the plasma concentrations and action of many anti-HIV drugs, particularly drug absorption. People who take certain antiretrovirals must therefore follow dietary restrictions, e.g., taking saqui-navir with a high-fat meal, or taking ddI (Videx) on an empty stomach. Different stages of pregnancy also affect plasma drug levels.

Adherence in particular is believed to play a major role both in maintaining drug levels (when adherence is maintained at least 95% of the time) and in triggering untoward treatment outcomes (when adherence is poor). For this reason, TDM is often used in other diseases to monitor adherence. However, if suboptimal HIV treatment outcomes result from wavering adherence to a drug regimen, measuring adherence with a variety of tools may be more appropriate than simply monitoring drug levels alone. For example, TDM might be combined with MEMS (medication event monitoring system, an electronic means of recording the time and date when a person takes each dose of medicine) and pharmacy refill data. TDM blood tests can show only that the most recent dose of drug has been taken, and even people who are not always adherent to their medications are likely to take their medication before being tested.

Several factors on the molecular level also may cause antiretroviral drug levels to fluctuate. For instance, a substantial number of PI molecules cling to proteins in the blood such as albumin or alpha-1 acid glycoprotein (AAG), which renders them unavailable for activity. Any drug dosing must take this phenomenon, known as protein binding, into account. Researchers recently have suspected that another protein, P-glycoprotein (P-gp), which transports drugs such as PIs and anticancer agents out of cells, may have a significant impact on the intracellular distribution and effectiveness of all antiretroviral drugs.

Drug-drug interactions must also be addressed in any evaluation of plasma drug levels. This is especially true of antiretrovirals, as both PIs and NNRTIs are metabolized by the cytochrome P450 system in the liver.

As a result, concentrations of certain HAART drugs can be dramatically increased or decreased when used in combination. This is also the case with many agents used to prevent or treat opportunistic infections (OIs) such as rifabutin (Mycobutin) and azole antifungals (e.g., fluconazole [Diflucan]). More studies need to address the issue of drug-drug interactions and toxicities in multidrug regimens as most drugs have been investigated only as monotherapy (i.e., not combined with other drugs).

Real-World Factors

Even if drug monitoring can be demonstrated to show utility in the clinic when treatment outcomes are suboptimal, the results generated by TDM assays must be interpreted by someone with sufficient expertise. The number of such experts may not be much higher than the undoubtedly small number of sites in the world presently offering TDM tests. Furthermore, these tests have yet to be standardized and are beset by quality-control problems that diminish their accuracy. In addition, there is no consensus on which specific measurement will prove to be most useful to clinicians. Possibilities include the Cmin (the trough, or minimum drug concentration), the Cmax (the maximum drug concentration), concentration ratios, or the AUC (area under the curve, a measure of total drug concentration in the plasma over a period of time [usually 12-24 hours]). Then there is the logistical problem of the assays themselves, which must be sent to special laboratories for analysis. Slow turnaround of TDM test results could make drug monitoring risky for someone experiencing treatment breakthrough as drug resistance may emerge between taking the plasma sample and acting on the interpreted lab results.

Another barrier to widespread use of TDM is the current high cost of the assays, approximately $150 to $200 per drug. Who will pay for these expensive diagnostic tests? Insurance companies will need proof of their utility before they agree to foot the bill. And pharmaceutical companies are not likely to warm to the idea of subsidizing TDM, even if it is shown that better treatment outcomes will ensure that people take their drugs for longer periods of time.

Will Customization Become Routine Practice?

The ATHENA Trial

TDM has rarely been assessed prospectively (forward in time). However, one ongoing prospective trial, known as ATHENA, has enrolled over 390 subjects in the Netherlands into one of two randomized arms. Those in the TDM arm have their drug levels periodically monitored; the resulting data are shared with their physicians, who have been instructed to alter doses if drug concentrations do not fall within 75-200% of the Cmin. Those in the other arm also have their drug levels monitored but the results are not revealed to their physicians.

An update report given by Dr. Hoetelmans and David Burger, PharmD, PhD, of St. Radboud University in Nijmegen, the Netherlands, suggested that ATHENA may fall short of its goal of providing conclusive evidence supporting TDM in the clinic. The report was presented at the First International Workshop on Clinical Pharmacology of HIV Therapy, held March 30-31, 2000, in Noordwijk, the Netherlands. Preliminary data from ATHENA showed that PI levels have been more than 200% above the Cmin in 5-11.5% of subjects, and less than 75% of the minimum effective concentration in 26-41% of subjects. Perhaps not surprisingly, some subjects in the TDM arm have already experienced treatment failure; Dr. Burger admitted that physicians simply may not be adjusting drug doses appropriately. He also surmised that some physicians in the control arm may be monitoring drug levels and adjusting doses on their own, which could turn ATHENA into a pointless exercise. More definitive results from the trial should be forthcoming in the next year.

The Promise of Better Therapy

Another factor might render TDM less useful, aside from the lack of conclusive trial data. PIs are now often coadministered with low-dose ritonavir (Norvir, another PI drug), with or without FDA approval (only the combination of lopinavir and ritonavir, known as Kaletra, has been approved). When currently marketed PIs are combined with small amounts of ritonavir they may be taken at lower-than-prescribed doses, since ritonavir causes other PIs to be metabolized (broken down) at a much slower rate in the liver. This produces highly desirable synergistic effects, namely an increased AUC of the principal PI in the combination and a trough plasma concentration significantly greater than the IC95 of wild-type, or nonmutated, HIV. (The IC95 is the drug concentration needed to inhibit viral replication by 95%, one of several markers of minimum drug efficacy.) Such potent double-PI combinations -- with lower drug doses often achieving more sustained viral suppression -- theoretically preempt the need for TDM. ATHENA will not shed any light on the implications of this clinical scenario since the trial began before dual-PI therapy became widespread.

The development of more powerful anti-HIV agents with an easier pill burden and fewer side effects may likewise diminish the appeal of TDM. In fact, some experts feel that the current interest in TDM deflects attention away from the need to develop these new-generation drugs. Doubtless, the energy and will to explore both are necessary.

Nicholas Cheonis is Associate Editor of BETA.

Selected Sources

Back, D.J. and others. Therapeutic drug monitoring of antiretrovirals: ready for the clinic? Journal of the International Association of Physicians in AIDS Care ( February 2000.

Barry, M.G. and others. Pharmacokinetics and potential interactions amongst anti-retroviral agents used to treat patients with HIV infection. Clinical Pharmacokinetics 36: 289-304. 1999.

Barry, M.G. and others. Variability in trough plasma saquinavir concentrations in HIV patients: a case for therapeutic drug monitoring. British Journal of Clinical Pharmacology 45(5): 501-502. May 1998.

Burger, D.M. and others. Low plasma concentrations of indinavir are related to virological treatment failure in HIV-1-infected patients on indinavir-containing triple therapy. Antiviral Therapy 3(4): 215-220. December 1998.

Emini, E.A. Resistance to anti-human immunodeficiency virus therapeutic agents. Advances in Experimental Medicine and Biology 390: 187-195. 1995.

Hoetelmans, R.M.W. and others. The effect of plasma drug concentrations on HIV-1 clearance rate during quadruple drug therapy. AIDS 12(11): 111-116. July 30, 1998.

Hugen, P. and others. Compliance to HIV-protease inhibitors (PIs) is more accurately measured by combining various methods. XIII International AIDS Conference. Durban, South Africa. July 9-14, 2000. Abstract ThPeB5029.

Mascolini, M. Is TDM too darn much? (and other prickly PK questions). Journal of the International Association of Physicians in AIDS Care ( May 2000.

Paterson, D.L. and others. Adherence to protease inhibitor therapy and outcomes in patients with HIV infection. Annals of Internal Medicine 133(1): 21-30. July 4, 2000.

Piscitelli, S.C. The limited value of therapeutic drug monitoring. Medscape, Inc. (
). 1999. (Click "Cancel" when prompted for a password, then register with Medscape to view article.)

Robbins, B.L. and others. Intracellular triphosphate concentrations of d4T and 3TC in HIV infected patients. 40th Interscience Conference on Antimicrobial Agents and Chemotherapy. Toronto. September 17-20, 2000. Abstract 1168.

Back to the SFAF BETA Autumn 2000 contents page.