Clinical Considerations in Therapeutic Drug Monitoring
Prior to ordering either a drug-level test or a phenotype for any patient, clinicians should evaluate the following
Both TDM and phenotyping are predicated on the assumption that the patient is taking all of his doses of all of his medications as prescribed. Poor adherence will skew the results of either test -- and, of course, poor adherence, rather than individual variations in pharmacokinetics, IC50, and viral type, accounts for a large number of cases of therapeutic failure. Therefore, a thorough appraisal of the patient's actual adherence -- as opposed to his self-reported compliance with therapy -- is a critical first step. Clear, candid, straightforward, non-judgmental discussions with patients about adherence-related issues may prevent unnecessarily switching of therapies and may improve the overall efficacy of a given regimen through simple modifications. Moreover, whenever such discussions uncover a discrepancy between actual adherence and self-reported compliance, they will prevent the clinician from ordering expensive and unnecessary tests. In all instances drug-level tests should be conducted only after a run-in period of at least 7 to 14 days, to ensure that all assigned doses have indeed been taken.
Potential Drug-Drug Interactions
All treatment failures, whether they involve increases in viral load or detectable viremia in patients who have previously had undetectable HIV RNA levels, should be carefully evaluated for drug-drug interactions. These inter-actions can lower serum drug levels and/or increase toxicities. One classic example of such insidious interactions is the one seen when efavirenz and indinavir are given concurrently, but drug-drug interactions are not limited to those that occur when antiretroviral agents are administered concurrently. The treatment of patients coinfected with HIV and TB is so fraught with the possibility of reduced plasma trough concentrations and heightened toxicities that clinicians are urged to consult experts on HIV-TB coinfection before prescribing rifamycins to patients taking NNRTIs or protease inhibitors. Moreover, certain antiretroviral agents, notably ritonavir, interact with the widely-used recreational drug Ecstasy, occasionally with fatal results. Therefore, clinicians must consider all of the drugs a patient is taking -- prescribed and over-the-counter, legal and illicit -- to be absolutely certain that drug-drug interactions can be ruled out as the cause of treatment failure.
Evaluation of the Entire Antiretroviral Regimen
We now have data showing correlations between the plasma levels of individual antiretroviral agents and response to therapy, but it is the potency and activity of the entire regimen, not just the most-studied agents, that should be evaluated. Numerous clinical studies have shown that antiretroviral regimens with the greater number of active agents are superior to regimens with fewer active components -- even when the latter regimens contain a single agent that is more active than the most active agent in the former regimens. Genotypic testing should be used to assess the efficacy of nucleoside analogs, whose active component, the intracellular triphosphate moiety, can be measured through TDM only with considerable difficulty.
Benefits Versus Cost
On average, plasma drug-level tests cost between $85 and $100 per sample, and most laboratories will only sample one drug level during each run of the assay. The time from sample submission to reporting of results can range from seven to 30 days. Commercial phenotype assays cost between $750 and $900 per sample. These assays take between two and three weeks to complete. Some, but by no means all, third-party payers will cover the cost of one or both of these tests. Although the combination of TDM and phenotypic testing does provides invaluable clinical information -- especially when it comes to choosing a successor therapy for an extensively pretreated patient -- that information comes at a price.
Table 2: Conversion Factors for ?M to ?g/ml*
|* Multiply ?M IC50 value by factor to determine patient's IC50 in ?g/ml.|
Once the results are reported, the clinician must take care to evaluate those laboratory data using the same units of measurement (e.g. ?M or ?g/mL) that the lab used. For example, our own lab reports drug levels in ?g/mL, while ViroLogic's assay reports the patient's IC50 as ?M (Table 2), so we must convert one or the other to establish equivalence. Finally, the clinician should determine what types of drug levels could be obtained for each of the drugs in the patient's regimen (Table 3), and then use that information to evaluate the patient's IC50 values for each of those agents.
Table 3: Plasma Levels of Protease Inhibitors
|Wild type IC50 (?g/ml)^||Trough level* (?g/ml) for mono-PI regimen||Trough level* (?g/ml) for dual-PI regimens with RIT|
600/100 -- 1.90##
900/100 -- 2.90##
1200/200 -- 2.18
400/400 -- 1.51#
800/100 -- 1.61
800/200 -- 3.71
400/400 -- 0.7
400/100 -- 5.5
400/200 -- 5.6
|ND -- Not done|
^ Upper end of IC50 values for reference standard using the PhenoSense? Assay
# -- Also reported as 0.54 ?g/ml in a different study
## -- Modeling data only
* It should be noted that the values reported in Table 3 represent mean values from clinical trials. Given the wide inter-patient variability seen with many of these regimens, a large proportion of patients will have levels significantly higher or lower than the mean values.
Using the clinical guidelines provided above, clinicians can apply the information derived from TDM and phenotypic testing in the following clinical situations
The chief clinical benefit of obtaining drug levels in this patient population is to identify patients with levels that fall toward the low end of the variability curve -- so that these patients can be shifted to the upper range of the curve through upward titration of their assigned doses. With timely intervention, these patients may be expected to experience more pronounced and rapid suppression of viral replication, and, potentially, a more durable antiretroviral response. Drug levels on this group of patients should be obtained at the first follow-up clinic visit, two to four weeks after the first viral load is obtained and in conjunction with the second HIV RNA measurement.
The practical benefits of obtaining a viral phenotype at this stage of infection are limited, and they may be outweighed by their cost. As a general rule, clinicians may prefer to reserve phenotypic testing of antiretroviral-na?ve patients for that subset of patients who are most likely to have been infected with multidrug-resistant viral strains (see "Multidrug-Resistant HIV and Primary Infection"). By and large, this subpopulation consists of the HIV-negative sexual partners of seropositive individuals on multidrug therapy and sexually promiscuous men who have sex with men.
The information obtained when the results of TDM and phenotyping are combined is likely to be of greatest clinical utility -- and benefit -- in this patient population. Virtually all of these individuals are veterans of successive, and successively less successful, antiretroviral regimens, and as a result they display varying levels of resistance (and cross-resistance) to a significant number of the agents currently approved to treat people with HIV infection. In this patient population -- a population with sharply limited therapeutic options, multiple codon mutations, and diminished life expectancy -- an evaluation of the individual patient's IC50 and plasma drug levels can provide potentially valuable treatment information.
In our practice, we have used phenotypic testing in this patient population to identify potential viability in agents that genotypic tests have identified as strongly resistant to the patients' viral isolates. This disparity between the information obtained from genotyping and that obtained from phenotyping has been described by others (32), and it may become even more pronounced as use of this technology increases and the technology itself evolves.
Although cross-resistance to the most potent class of antiretroviral agents, the protease inhibitors, is often regarded as the greatest stumbling block to developing an effective salvage therapy for heavily pretreated patients, our own experience has been that it is easier to find a combination of protease inhibitors that will overcome viral resistance than it is to find an effective nucleoside analog "backbone" for that salvage regimen. Although cross-resistance among the nucleoside analogs is certainly a factor, the toxicities associated with the chronic administration of these agents -- and the resistance to taking these drugs that such adverse events engender in patients -- are also factors (see "Managing the Adverse Effects of Antiretroviral Therapy").
Andrew D. Luber, Pharm.D., is Executive Director of Pacific Oaks Research, Beverly Hills, CA.
W. David Hardy, M.D., is Associate Clinical Professor of Medicine, UCLA School of Medicine, Los Angeles, CA.