The utility of therapeutic drug monitoring (TDM) is still controversial. While common in Europe, it has not yet become part of standard clinical practice in the United States mostly because no studies have shown a definitive benefit to TDM that justifies the added costs.

The main idea behind TDM is that the standard dose of an antiretroviral medication may not be right for all patients. Some patients may be getting too little of the medication into their bloodstream, which puts them at high risk for virologic failure and resistance. Other patients may be getting too much medication, which can lead to side effects and toxicities. By measuring the level of drugs in the blood, healthcare providers can make recommendations to increase (or decrease) doses so that blood levels can be brought into the appropriate range. Patients who may particularly benefit from TDM include those who are very thin, very heavy, pregnant or coinfected with hepatitis B or C.

In a study presented at this conference by Richard Haubrich of the University of California-San Diego, researchers from the California Collaborative Treatment Group enrolled 199 patients in a TDM trial. To be eligible for this trial, patients needed to be starting a new protease inhibitor (PI)- or non-nucleoside reverse transcriptase inhibitor (NNRTI)-based antiretroviral regimen. Other mean baseline patient characteristics included the following:

Baseline Patient Characteristic	Mean Value
CD4+ cell count	189 cells/mm3
HIV RNA	5.2 log10 copies/mL
Age	40 years
Weight	164 lbs
Proportion starting first antiretroviral regimen	29%

After 2 weeks, the patients' PI and/or NNRTI blood levels were measured. A panel of experts then reviewed those drug levels, along with the patients' treatment history, CD4+ cell count, plasma HIV-RNA level and clinical toxicities, with the goal of arriving at a recommendation to either change the PI/NNRTI dose or to leave it the same. A recommendation was issued for all study patients, but the patients were randomized into 2 groups: In one, the investigators received the experts' recommendations, while in the other, no recommendation was received, thus guaranteeing that no dose change would occur.

Overall, the expert panel recommended changing the PI or NNRTI dose in 67 patients (38%). All but 3 recommendations were to increase the dose. The factors associated with needing to increase the dose were: higher weight or body mass index (i.e., weight adjusted for height), use of either efavirenz (EFV, Sustiva, Stocrin) or lopinavir (LPV), and non-Hispanic ethnicity. In the adjusted analysis, weight and use of either efavirenz or lopinavir remained associated with a dose adjustment recommendation. Interestingly, adherence, age and gender were not associated.

The main result of this study -- whether a dose adjustment actually improves virologic outcome -- has not yet been analyzed. However, these results are important because they show that a significant number of patients taking antiretroviral medications will have inappropriately low (or, less commonly, high) drug levels. If we can identify those patients who are at greatest risk, we can target TDM for those patients who are most likely to benefit from it. Admittedly, these results also show that not all drugs may need monitoring, and it also appears that we cannot yet accurately predict which patients will need TDM.

In the poster discussion session during which this study was presented, a few key issues about TDM were raised:

1. There does not seem to be a consensus about what constitutes the "therapeutic range" for each antiretroviral drug. To complicate matters further, the range is probably different for patients who have drug resistance, since higher drug levels may be necessary to control HIV in these patients.

2. TDM for nucleoside/nucleotide reverse transcriptase inhibitors (NRTIs) does not seem practical at this point because it is the drug levels inside the cells, not the blood levels, that really matter and measuring these levels is quite complicated.

3. Drug levels can vary widely within a patient, making TDM more complicated. In a presentation by Richard Nettles et al, Nettles showed that the drug levels in a given individual may vary by as much as 43% for PIs and 26% for NNRTIs.¹ This indicates that it may be necessary to gauge drug levels multiple times before making any treatment decisions.

Intriguingly, another abstract, by Monica Gandhi et al, measured drug levels in a completely different manner from that used in either the Haubrich or Nettles studies: Gandhi et al used a sample of a patient's hair to estimate exposure to PIs over the prior 6 months.² Patients with lower levels of PIs in their hair were found to be less likely to respond to treatment. However, although this test would be useful in research settings to assess the relationship of drug levels to treatment response, it may be difficult to use in clinical practice to guide patient management because TDM usually involves measuring a blood level and making a dose change within weeks after starting a new antiretroviral medication, not 6 months later.

Despite this newly presented research, the utility of TDM has still not been definitively proven. However, a healthcare provider in the United States can consider TDM if he or she has access to it, has a plan for what to do with the results and has a specific reason for ordering the test. Fortunately, TDM is a focus of ongoing studies, including studies like the one discussed above by Haubrich et al, which should help clarify this important issue in the future.

Footnotes

1. Nettles R, Kieffer T, Parsons T, et al. Frequent sampling in virologically suppressed patients taking HIV protease inhibitors or non-nucleoside reverse transcriptase inhibitors defines intra-individual pharmacokinetic variability. In: Program and abstracts of the 12th Conference on Retroviruses and Opportunistic Infections; February 22-25, 2005; Boston, Mass. Abstract 642.

2. Gandhi M, Aweeka F, Jayawardene A, et al. Determination of ART exposure using hair: a measure of average exposure highly correlated with treatment outcome. In: Program and abstracts of the 12th Conference on Retroviruses and Opportunistic Infections; February 22-25, 2005; Boston, Mass. Abstract 744.