icon_folder.gif   Conference Reports for NATAP  
  ICAAC 41st Interscience Conference on Antimicrobial Agents and Chemotherapy
Chicago, Illinois, December 16-19
Back grey_arrow_rt.gif
Pharmacokinetic Issues of Antiretrovirals (Part 2)<
Written by Stephen Piscitelli, Pharm.D. Associate Director, Clinical Pharmacology Tibotec-Virco
Tenofovir PK
BMS-232632 (Atazanavir) PK
Efavirenz: blood levels & side effects
Indinavir, nelfinavir: blood levels correlate with effect
Pharmacokinetic studies of new antiretrovirals have always had a forum at the yearly ICAAC meeting. Several interesting studies were presented both for approved and investigational drugs. A number of papers also evaluated retrospective analyses of relationships between plasma concentrations and either virologic efficacy or adverse effects. Such data support the concept of therapeutic drug monitoring, however, correlations between blood levels and effects are only one part of successful TDM. For a more in depth summary of TDM, readers can refer to my summary on this website of this years Retrovirus meeting.
Tenofovir Pharmacokinetics
Scientists from Gilead reported on the pharmacokinetics of tenofovir DF (TDF) in two studies of HIV+ positive patients (n=17) and one trial in healthy volunteers (N=36).1 They statistically examined the pharmacokinetics of TDF looking at a variety of demographic factors. Overall, they found no difference in TDF pharmacokinetics with respect to HIV status, gender, age, or body weight. While this paper does not seem noteworthy, it is quite interesting that pharmacokinetic parameters did not differ between healthy volunteers and HIV-infected subjects. This concept does not appear to be true for a number of other antivirals including saquinavir, indinavir, and atovaquone as well as perhaps other PIs. This has been a constant debate among pharmacologists as to whether healthy volunteer data can be used to make predictions and dosage recommendations in HIV+ positive patients. These data support that TDF healthy volunteer pharmacokinetic data are valuable in guiding other studies and uses.
Pharmacodynamics of BMS-232632
Relationships between the pharmacokinetics of the investigational protease inhibitor, BMS-232632, and both efficacy and toxicity were presented in a poster by BMS scientists.2 Data from a Phase II study that included both the 400 mg and 500 mg once daily dose were evaluated to help guide in dose selection for Phase III trials. Investigators demonstrated that a higher estimated AUC (based on a limited number of blood samples) was a significant predictor of achieving a 1.5 log reduction and also a significant predictor of a bilirubin elevation > 2.5 gm/dl. Statistical evaluations looked at the probability of achieving a 1.5 log viral log drop and an increased bilirubin. Although the mean probability of viral load drop was lower with the 400 mg dose vs 500 mg (0.78 vs 0.9), they suggested the 400 mg dose should be selected since it still gave good virologic response and had a lower probability of increased bilirubin (0.17 vs 0.33). This is a well-performed analysis and demonstrates that efficacy must be balanced against toxicity and tolerability. Higher AUCs appear to be a good predictor of response, however, it is difficult and not practical to measure AUCs. Trough values are much easier to obtain although no information on the predictive ability of troughs was presented.
Efavirenz and CNS Side Effects
Four different papers examined efavirenz (EFV) concentrations and relationships with CNS side effects. Results were widely variable between studies ranging from significant relationships to a lack of any correlation. The first study by Nunez and colleagues from Spain compared 15 patients with Grade3 or 4 CNS toxicity to a control group of 36 patients without side effects.3 Patients with side effects had higher levels (4.9 vs 3.7 ug/ml). In a multivariate analysis, plasma concentrations were an independent predictor of outcome. Patients with a level of > 3.5 ug/ml had a significantly higher incidence of CNS side effects. While retrospective analyses such as these have limitations in the selection of patients, matching criteria, etc, these data are similar to a study reported last year that demonstrated a concentration > 4.0 ug/ml was associated with CNS effects. Debate continues on to whether NNRTIs might be candidates for TDM. These two studies combined suggest that higher concentrations of EFV may lead to increased toxicity. Conversely, DuPont scientists presented a population pharmacokinetic-pharmacodynamic analysis in 564 subjects in which they estimated the AUC, Cmax, and Cmin based on PK data collected in clinical trials.4 They noted no difference in EFV pharmacokinetics in those patients with moderate or severe CNS toxicity compared to those with no CNS symptoms. While this study had a much larger population, many of the levels were estimated based on PK modeling and were not actually observed concentrations. It is clear that retrospective analyses can show highly variable results depending on the patients selected, the degree of monitoring, level of adherence, predisposing conditions and other factors. This issue will require a prospective evaluation of NNRTI TDM before a definitive recommendation can be made on the usefulness of TDM in all patients taking EFV.
Hawkins and colleagues evaluated self-report surveys of adverse effects in 68 patients receiving EFV and 71 patients receiving PIs.5 Data were evaluated at week 4, week 24, and week 48. The authors state there was no difference in symptom scores in the primary analysis, but a subgroup analysis of those with less than the median duration of therapy demonstrated higher reported CNS effects compared to PIs, even out to 48 weeks. The authors╣ conclusions stated that CNS symptoms may not decline to the level seen in PIs for 1 year. This study has a number of limitations including the self-reported side effects, selection of patients, and especially a statistical subgroup analysis of which always raise a "red flag". This study is extremely difficult to interpret and clearly requires verification. Contrary to these data, another study showed that EFV CNS effects decrease over time and quality of life increases with duration of therapy.6 In addition, adherence remained high (>90%) with EFV-based regimens. From a practical standpoint, CNS side effects can be problematic and persistent for some patients but commonly do decrease with time. These studies, while appearing to be contradictory, enforce this message that we have already learned with extensive experience with the drug.
Protease Inhibitor Levels: Correlation with Effect-indinavir, nelfinavir
Four posters described retrospective analyses on relationships between levels of PIs and efficacy or toxicity. Investigators from Italy used population pharmacokinetics to estimate the AUC, Cmax, and Cmin of indinavir (IDV) and used these values to examine relationships with virologic outcome and toxicity.7 They observed that patients with an AUC above 25 mg/L*h had a significantly better decline in viral load than those below this breakpoint value. They also reported that renal toxicity (0.5 mg/dl increase above baseline) was more frequent in patients with a higher Cmax value. These data are similar to work presented by David Burger at the 2001 Retrovirus Conference. Although intriguing, patients in this study received IDV 800 mg q8h as a sole PI. Since IDV use is becoming increasingly common when given with low dose ritonavir (RTV), it is hard to interpret these data and extrapolate the results to patients on a boosted regimen. In addition, at least 2 blood samples and possibly three would need to be collected from patients to get a good estimate of AUC, which will likely be difficult in many clinic situations.
Investigators from France examined plasma levels of nelfinavir (NFV), its active metabolite (M8), and IDV at month 4 into therapy and then calculated how these levels may change over time and their relationship with outcome.8 They calculated an observed/predicted ratio between individual and mean time-adjusted population plasma drug levels. They reported that those patients with and O/P ratio of < 0.8 for NFV+M8 had a significantly higher risk of viral rebound (41%) vs those above 0.8 (27%). No relationships were noted for IDV. The investigators suggested that patients who have adequate NFV levels early on and achieve a response may not maintain these plasma levels over time and TDM should be considered. There have been discussions of exactly what happens to PI levels over a prolonged period. A decrease in levels with time has been noted for SQV but there are not many studies that address this important issue. Relaxing of food restrictions, addition of concomitant meds, and changes in adherence are a few possible reasons for changes in levels over time. These data are suggestive but it is not clear if M8 levels can be added to parent NFV levels in a 1:1 manner such as performed here. It depends on the IC50 of the metabolite which has not been well described. Still, this study demonstrates an important point that measuring plasma levels only once early into treatment may not be optimal for monitoring.
Two other studies examine NFV concentrations and outcomes. The first included 154 patients on NFV regimens (BID or TID) and examined drug levels along with resistance mutations for relationships with outcome.9 In patients with > 4 reverse transcriptase + protease mutations at baseline, virologic success was associated with higher a higher mean trough value (2 mg/L) versus failure (0.7 mg/L). The second study was presented by the manufacturer of NFV (Agouron) and looked at the effect of diarrhea on NFV absorption and outcome using data from a Phase III trial.10 Pre-dose and 2 hour post dose levels were drawn in 99 patients. Drug levels and response at 24 weeks were not significantly different in patients with and without diarrhea. M8 concentrations were not reported. In this study 72 patients had Grade 1 or 2 diarrhea and only 3 patients had Grade 3-4. While it would be interesting to see the effect of more sever diarrhea, many of these patients may choose not to continue therapy making the analysis less clinically relevant.
There have now been over 45 studies that have retrospectively examined relationships between antiretroviral drug levels and either toxicity or efficacy. These studies suggest that achieving higher drug levels is an independent predictor of virologic success. However, while this makes a compelling case for TDM, a number of ongoing prospective trials will better address this important issue.
1. Kearny BP, Gill SC, Flaherty JF, et al. Effect of demographic variables on the pharmacokinetics of tenofovir DF in HIV-infected patients and healthy subjects. 41st Interscience Conference on Antimicrobial Agents and Chemotherapy. Chicago, IL, September 22-25, 2001, abstract A-504.
2. Omara E, Cirincione B, Mummaneni V, et al. Population pharmacodynamic assessment of the safety and antiviral activity of BMS-232632. 41st Interscience Conference on Antimicrobial Agents and Chemotherapy. Chicago, IL, September 22-25, 2001, abstract A-507.
3. Nunez M, Gonzalez de Requena D, Gallego L, et al. Impact of efavirenz plasma levels on the development of severe central nervous system toxicity. 41st Interscience Conference on Antimicrobial Agents and Chemotherapy. Chicago, IL, September 22-25, 2001, abstract I-1724.
4. Hawkins T, Geist C, Young B, et al. Efavirenz induced CNS side effects persist for one year compared to protease inhibitors. 41st Interscience Conference on Antimicrobial Agents and Chemotherapy. Chicago, IL, September 22-25, 2001, abstract I-1725.
5. Fumaz CR, Tuldra A, Ferrer M, et al. Impact of central nervous system disorders on quality of life and adherence to treatment of HIV-1+ patients treated with efavirenz. 41st Interscience Conference on Antimicrobial Agents and Chemotherapy. Chicago, IL, September 22-25, 2001, abstract I-1726.
6. Fiske WD, Joshi AS, Labriola DF. An assessment of population pharmacokinetic parameters of efavirenz on nervous system symptoms and suppression of HIV-RNA. 41st Interscience Conference on Antimicrobial Agents and Chemotherapy. Chicago, IL, September 22-25, 2001, abstract A-1727.
7. Gatti G, Di Biagio A, De Pascalis CR, et al. The relationship between systemic exposure to indinavir and response (virologic efficacy and renal toxicity). 41st Interscience Conference on Antimicrobial Agents and Chemotherapy. Chicago, IL, September 22-25, 2001, abstract A-1732.
8. Le Moing V, Peytavin G, Ecobichon JL, et al. Plasma levels of indinavir and nelfinavir at time of viral response may have a different impact on the risk of further viral failure in HIV-infected patients. 41st Interscience Conference on Antimicrobial Agents and Chemotherapy. Chicago, IL, September 22-25, 2001, abstract A-1733.
9. Pellegrin I, Brelh D, Garrigue I, et al. Virologic response to nelfinavir-containing regimens: individual pharmacokinetic parameters and drug resistance mutations. 41st Interscience Conference on Antimicrobial Agents and Chemotherapy. Chicago, IL, September 22-25, 2001, abstract A-1734. 10. Hsyu PH, Flexner C, Chu A, et al. Correlation of efficacy, nelfinavir pharmacokinetics and diarrhea in treatment na´ve HIV positive patients receiving nelfinavir, zidovudine and lamivudine. 41st Interscience Conference on Antimicrobial Agents and Chemotherapy. Chicago, IL, September 22-25, 2001, abstract I-1735.