8th Annual Retrovirus Conference



Section by Stephen C. Piscitelli, PharmD, with selected contributions by Harvey S. Bartnof, MD and Jules Levin

“Therapeutic Drug Monitoring”
Therapeutic drug monitoring (TDM) is individualizing for each patient his/her drug doses based upon measuring blood plasma concentrations. Clinicians have used these principles for years to adjust doses of drugs including theophylline (asthma drug), aminoglycosides (antibiotics), digoxin (heart medication), and anticonvulsants (for seizures or ‘fits’). There is growing evidence that TDM may be useful to ensure that HIV-infected patients have adequate blood levels or that their plasma concentrations are not too high that toxicity develops. While TDM of anti-HIV drugs has been used commonly in Europe, it has not yet been well studied or utilized in the US. However, at this year’s Retrovirus Conference, an entire poster session was devoted to the topic, and TDM issues also were presented as part of oral presentations on "pharmacology" (study of drugs) of HIV drugs. To read more about those studies, see the NATAP website Retrovirus Conference coverage: (

Issues that are in the forefront in the area of TDM research include:

  1. Will TDM help in the management of HIV positive patients taking HAART?

  2. Is TDM cost-effective?

  3. Will TDM be helpful in preventing or managing anti-HIV drug toxicities?

  4. What is the best blood level to use in TDM: lowest ("trough"), maximal or random applied to "total exposure" ("area-under-the-curve") concentration?

  5. Is just one TDM sample enough? How many are necessary?

  6. Is TDM beneficial for treatment-naïve (never took) and/or treatment-experienced patients?

  7. Which anti-HIV drug classes should be used in TDM? NRTIs need to be evaluated intra-cellularly & there is no commercial test for this yet.

  8. PharmAdapt" evaluated TDM but was not designed properly (abstract 260B)

  9. How important is the wide inter- and intra-patient variability in TDM?

  10. What is the best way to incorporate drug resistance information with TDM?

  11. Are drug resistance tests and TDM best used by any of the following proposed calculations: "In Vivo Potency Index" (abstract 732); "Virtual Inhibitory Quotient" (abstract 523); a regular "Inhibitory Quotient" (abstract 523); or the "Virtual Phenotype?" And do any of these calculations need to be adjusted for "protein binding?"

  12. What is the optimal "turnaround" time for TDM results and is there a delay after which the results are no longer meaningful?

Unfortunately, at the present time, there are more questions than answers—research is ongoing. But studies are being planned to try and answer these questions. In addition to the abstract listed among the questions above, others that were useful included abstracts 259, 730, and 738).

Drug Level + Resistance + IC50 = virtual INHIBITORY QUOTIENT
A major limitation of using only the drug level to guide therapy (other than it does not describe long-term adherence) is that the optimal drug level may vary from patient to patient. Indeed, the patient with highly drug-resistant HIV will require higher drug levels than a patient naïve to (never took) therapy. Thus, successful attempts at TDM should also take into account a measure of the virus's sensitivity (resistance) along with the drug level. The use of the "IC50" (inhibitory concentration that inhibits 50% of HIV in the laboratory) from phenotype drug resistance testing could theoretically provide an optimal target concentration in the patient with highly drug-resistant virus. Two research groups reported using drug concentration data with the "protein binding-corrected" IC50 value. The first study, reported by Courtney Fletcher, PharmD of the University of Minnesota examined 8 treatment-naïve patients receiving AZT/3TC/IDV (abstract 732). He calculated something he called the "in vivo potency index" (IVP). This value was determined for each drug by dividing the drug level in blood by the virus’ susceptibility or resistance and then adding up the values of all of the drugs in a regimen together. For example, the authors calculated the ratio of free IDV (not bound to plasma proteins) or of "intracellular triphosphate" concentrations of the NRTI drugs (active forms) to the "IC90" of the virus at baseline. When the ratios were summed for the 3 drugs, patients who had a summed ratio greater than the median (midpoint, half above, half below) had a faster time to achieving undetectable plasma HIV-1 RNA levels than those below the median. Four of 4 patients with a higher median ratio achieved an undetectable viral load by week 24, compared to only 2 of 4 patients with a summed ratio less than the median.

There has been much talk recently about the use of drug "Inhibitory Quotients" (IQ’s) as a better measure of anti-HIV drug resistance. Dale Kempf, PhD of Abbott Laboratories noted in his talk that the IQ concept comes from the microbiology literature (abstract 523). It refers to a measure that includes information about a drug’s actual or expected concentration in blood (e.g., plasma "Cmin" or minimum concentration) divided by a measure of the pathogen’s ("germ’s") susceptibility (resistance) to that drug (e.g., IC50 or IC90, meaning "Inhibitory Concentration" of drug that would inhibit 50% or 90% of the pathogen’s growth in the laboratory). So the "inhibitory quotient" could be written: IQ = Cmin/ IC50. Theoretically the IQ is a better measure of resistance, because, as we know for most drugs, resistance is relative to a drug’s concentration. Investigators at Abbott showed last October at the Glasgow meeting that IQ predicts response to lopinavir/ritonavir (Kaletra, dual PI drug formulation) combination therapy.

“Virtual Inhibitory Quotient” (vIQ)
Now, Dr. Kempf reported on the predictive capacity of a "Virtual Inhibitory Quotient" (vIQ) in predicting response to ritonavir "intensification" of patients taking indinavir (IDV, Crixivan, PI drug)-based HAART and who had ongoing viremia (at least 50 copies per milliliter) (abstract 523). (Patients "switched" to IDV/RTV, 400 mg of each twice-daily.) The Virtual IQ uses the "Virtual Phenotype" (vPT, see page 25) in place of the actual measured IC50, and the actual "Cmin’s" for IDV were measured in a pharmacokinetic (PK, drug concentration) study of 37 patients. Therefore, the "vIQ = Cmin / vPT x IC50"for "wild type" (no mutations) HIV that is corrected for the presence of 50% human serum to approximate better what takes place in people. (Note that the vPT is given as average fold-change in IC50). This is similar to the IVP (In Vivo Potency Index, see above), except that it corrects for "protein-binding" effects instead of actually measuring free drug. Dr. Kempf’s analysis indicated that not only did the vIQ for indinavir predict virologic response over 48 weeks but it appeared to discriminate response from lack of response better than did a simple "genotype score" (number of PI drug-related resistant mutations) or the vPT. For patients with a vIQ greater than 2, the viral undetectability rate after 3 weeks was 90% (18 of 20) and after 24 weeks was 65% (11 of 17). By contrast, for patients with a vIQ less than 2, the 3-week response rate was 13% (1 of 8) and the 24-week response was zero (0 of 8). The potential benefits of the Virtual Inhibitory Quotient will need to be defined for other drugs and in more studies.

Limitations of these 2 studies were that they were performed retrospectively and in a small number of patients. However, they demonstrate an interesting concept that integration of the phenotype and the drug level may be used to optimize and guide therapy. Much research will still be required to validate the Cmin/IC50 ratio and breakpoints for this ratio will need to be identified. Also, there needs to be consensus on how to best correct for protein binding. With highly bound drugs like lopinavir and saquinavir, it is very difficult to measure free drug levels. A number of correction factors can be used but there remains much disagreement on which is best. Another question is whether the trough (lowest) level or the "AUC" (area-under-the-curve or total drug exposure) concentration is the most important. Still, this appears to be exciting research and may be a tool for monitoring patients in the future.

Drug Transport Proteins
Another area of increasing interest is drug transport proteins. The term "drug transporter" refers to a group of proteins that affect how drugs get into, and out of cells. Some drug transporters appear to pump HIV drugs out of cells, or inhibit the absorption of HIV drugs in the GI (gastrointestinal, stomach-intestine) tract, thereby potentially decreasing the bioavailability of anti-HIV therapy. Unfortunately, this field is still relatively in its infancy and most of the work has only been performed in the laboratory. But it does raise some important questions that may help explain the wide variability in virologic response to HAART between patients (abstracts 260, 737). Drug transport proteins that were discussed at the Drug Transporter Symposium include "P-glycoprotein," a cellular "efflux pump" (pushes drugs out of cells) encoded by the "MDR1" (multi-drug resistance) gene (abstracts S1, S3, S4) and multidrug resistance protein (MRP) 4 (abstract S2). To read more about this information, see the NATAP website:

Drug-Drug Interactions
Three new drug interaction studies with the antibiotics rifabutin or rifampin were presented (abstracts 32, 741, 742). To read about these reports, see the website

Complementary and alternative medicines are widely used in HIV-infected patients despite a general lack of knowledge on their adverse effects and drug interactions. Increased attention has recently been paid to these therapies (abstract 497) and two posters at the 8th CROI evaluated herb-drug interactions.

The effect of garlic capsule supplements on the drug levels of saquinavir (SQV, Fortovase, PI drug) was studied in 10 healthy volunteers by Stephen C. Piscitelli, PharmD of the NIAID (abstract 743). Subjects had blood plasma samples collected for SQV levels while taking it alone (standard dosing) and then after 3 weeks of a twice-daily garlic supplement. SQV levels were again studied alone after a 3-week "washout" (no garlic) period. Overall, there was a decrease in the SQV "AUC" (area-under-the-curve or total exposure) concentration and "Cmin" (minimum concentration) of approximately 50%. The mechanism appeared to be a prolonged "induction of metabolism," since most patients did not return back to their baseline levels even after the 3-week washout period. Although these data are interesting, this study introduces more questions than it answers. For example, would a decrease be seen when SQV is given with ritonavir, as commonly used? What is the actual mechanism, and does dietary garlic (not capsules) affect anti-HIV drug levels? Ongoing studies with other herbs may help to answer some of these issues. The NIH is currently conducted a study with Milk Thistle. Also, the results might be different in HIV positive persons. Until more is known, it would seem prudent for persons taking a PI drug not to take garlic supplement capsules.

Marijuana is often used in HIV positive persons for a variety of reasons, but it is currently unknown if its use affects the "pharmacokinetics" (PK, drug levels) of PI drugs. Bradley Kosel, PharmD of San Francisco General Hospital (SFGH) in California presented the results of a study that examined the effects of marijuana on the "PK" of indinavir (IDV, Crixivan) or nelfinavir (NFV, Viracept) in HIV-infected patients (abstract 745). Inpatient study subjects were randomized to marijuana 4% "THC" (active marijuana component) cigarettes that were smoked, dronabinol capsules (Marinol, with THC) taken orally, or placebo, all 3-times daily before meals. The 67 patients (96% men, 52% non-White) were taking stable IDV- or NFV-based HAART. Blood plasma samples were collected at baseline and after 14 days. The results showed that the patients receiving the cigarettes had non-significant decreases in IDV and NFV "Cmin" of 34% and 12%, respectively. The only statistically significant change was a modest 14% decrease in the IDV "Cmax" (maximal concentration). Measurements of "M8," the active metabolite (by-product) of NFV, were not reported. The other two treatment groups did not have any significantly different levels of IDV or NFV PK measurements. Of note, there were wide variabilities in patient measurements. Based upon these results, it is unlikely that marijuana cigarettes will markedly affect PI drug concentrations. In addition, Donald Abrams, MD also of SFGH reported that there were no significant changes observed in CD4 counts or HIV viral loads among the patients (abstract 744), although changes in these parameters may take longer to occur.

< All newsletters back to index >