 |
|
|
| |
CLINICAL PHARMACOLOGY AT THE 14TH CROI
|
| |
| |
February 25-28, 2007
Los Angeles Convention Center, CA
John G. Gerber, MD
Jennifer J. Kiser, Pharm D
Abbreviations used in this review:
3TC=lamivudine, FTC=emtricitabine, ZDV=zidovudine, TDF=tenofovir disoproxil fumarate, TFV=tenofovir, d4T=stavudine, ABC=abacavir, EFV=efavirenz, NVP=nevirapine, DRV=darunavir, IDV=indinavir, RTV=ritonavir, SQV=saquinavir, SQV/r=saquinavir/ritonavir, SQV-HGC=saquinavir hard gel capsule, ATV=atazanavir, ATV/r=atazanavir/ritonavir, LPV=lopinavir, LPV/r=lopinavir/ritonavir, NFV=nelfinavir, APV=amprenavir, fAPV=fosamprenavir, PI=protease inhibitor, NNRTI=non-nucleoside reverse transcriptase inhibitor, N(t)RTI=nucleos(t)ide reverse transcriptase inhibitors, ARV=antiretroviral, HAART-highly active antiretroviral therapy, PK=pharmacokinetics, QD=once daily, BID=twice daily, TDM=therapeutic drug monitoring, AUC=area under the concentration time curve, Cmax=maximum concentration or peak, Cmin=minimum concentration or trough
Introduction:
The 14th CROI was held at the downtown Los Angeles Convention Center. Although the Convention Center was certainly adequate for this meeting, there was a long distance between the hotels and the Convention Center so commuting was time consuming.
This year there were too few Clinical Pharmacology oral presentations to make an impact at this meeting. There were several posters dealing with important pharmacologic issues of HIV therapy, but TDM abstracts were few and far between. As stated in last year's report, the use of newer and pharmacologically improved ARV drugs will likely make TDM a less critical issue of ARV therapy.
In organizing this report, we will review the abstracts that we felt were clinically important. No doubt we have missed some important presentations and we would like to apologize up front for that. In addition the opinions expressed in this report are clearly ours based on our experience in Clinical Pharmacology and HIV Medicine. We will organize this report by specific Clinical Pharmacology topics except for the three oral presentations and Dr. Angela Kashuba's invited lecture which will be reviewed in a separate section. In addition, the oral abstract on the effect of entecavir on HIV replication will also be reviewed separately. We felt that the clinical implications of those data required attention as entecavir was FDA approved several years ago for the treatment of hepatitis B virus (HBV) infection with the understanding that the drug did not have any anti-HIV activity. It was assumed that the drug could be safely administered to HBV-HIV co-infected subjects without any concern for the development of HIV resistance mutations.
Entecavir inhibits HIV replication (abstract 136LB)
This abstract presented unexpected important data on the anti-HIV effect of entecavir. Entecavir is a cyclopentyl guanosine analogue with an EC50 ∼ 3.75 nM against numerous HBV strains. The drug at slightly higher concentrations also has activity against 3TC-resistant HBV. The drug was approved in 2005 for the treatment of HBV infection based on excellent efficacy and safety data in HBV infected patients. In vitro the drug had not demonstrated any activity against HIV and thus it became a preferred agent in the treatment of HIV-HBV co-infected subjects who did not require therapy for their HIV infection. However, Dr. Chloe Thio presented compelling data to suggest that entecavir does have activity against HIV and that it selects for M184V mutations on the reverse transcriptase which results in resistance to entecavir. These data have resulted in a Dear Doctor letter warning physicians about the danger of using entecavir in HIV/HBV co-infected subjects not yet on HAART. The impetus to study the effect of entecavir against HIV came from clinical observations in two HIV/HBV co-infected subjects who had a 1 log decrease in HIV-RNA on entecavir monotherapy. An additional subject was found to accumulate the M184V mutation on HIV reverse transcriptase on entecavir monotherapy. These investigators went to the laboratory and generated in vitro data showing that entecavir inhibited HIV infection with an IC50 between 0.1-1.0 nM (BristolMyersSquibb using a different assay reported an IC50 in the micromolar range, a concentration far above what is achieved in vivo). In addition, HIV with M184V demonstrated resistance to entecavir. These combined data strongly suggested that entecavir has anti-HIV activity. These findings are important because HIV-HBV co-infection is commonly observed in Western countries since both viruses can be sexually and blood transmitted. Although the development of M184V on the reverse transcriptase may not be the most problematic mutation that develops to ARV drugs, most initial regimens contain either FTC or 3TC and M184V would antagonize the antiviral effect of those drugs. We believe that more data in co-infected subjects on entecavir monotherapy need to be collected, but until those data are available, entecavir cannot be recommended for use in the treatment of HBV infection in subjects with concomitant HIV infection unless the HIV infection is effectively suppressed.
ORAL PRESENTATIONS (Session 18: Novel Approaches for Pharmacokinetic Assessment).
This session began with an elegant presentation by Angela Kashuba reviewing the use of drug cocktails to determine specific drug actions on CYP isozymes and transporters, known as phenotyping. She reviewed the way these cocktails are used and which drug specifically measures the activity of which CYP isozyme, P-glycoprotein (Pgp) transporter, and glucuronyl transferase. These probe drugs have been used successfully in identifying the effect of ARV drugs on specific CYP activity. She provided an example where an interaction between LPV/r and phenytoin needed an explanation. Dr. Kashuba's group has published data showing that when these drugs are co-administered the concentrations of both LPV and phenytoin are decreased (Lim et al., JAIDS 36:1034, 2004). This was surprising because RTV was thought to be an inhibitor of CYP2C9 and 2C19 and phenytoin uses these CYPs for metabolism. Using a phenotyping cocktail, they were able to demonstrate that the usual doses of LPV/r result in an induction of CYP2C9 and 2C19 (Yeh et al., JAIDS 42:52, 2006). In contrast, the phenotyping data on the effect of first dose and steady state TPV/r gave results that we found difficult to reconcile as to why TPV/r markedly reduces the plasma concentrations of SQV, LPV, and APV (poster 563, Vourvahis et al.). The drug cocktail determined that after the first dose of TPV/r hepatic and intestinal CYP3A4/5 are markedly inhibited and at steady state this inhibition is attenuated, but still present. In terms of Pgp activity, the intestinal activity is inhibited during the first dose of TPV/r but this inhibition is lost at steady state. Surprisingly, significant Pgp induction was not demonstrated. Maybe the difference in activity of CYP3A4/5 and Pgp secondary to TPV/r as compared to only RTV is enough to explain the decline in HIV-PI concentrations, but the 90% reduction in AUC of SQV when co-administered with TPV/r would require a more potent induction in Pgp if this is the main transporter of SQV out of the mucosa. It is important to consider that other intestinal transporters that accelerate the exit of PIs from intestinal mucosa are induced by TPV/r. Dr. Kashuba concluded that it is necessary to develop high affinity and safe probe drugs for many of the important drug transporters in the gastrointestinal tract and systemic locations in order to phenotype accurately the effect on ARV drugs and predict important drug-drug interactions.
Of the three oral presentations that followed Dr. Kashuba's review, two were of excellent quality with significant clinical implications and one kept us puzzled as to how the presented information could be used in clinical practice.
Monica Gandhi presented data on the hair concentrations of LPV and RTV in HIV-infected women on at least 6 months of LPV/r from the Women's Interagency HIV Study (WIHS) group (abstract 51). Their a priori hypothesis was that hair concentrations of PIs may indicate an average level of exposure of these drugs over months and therefore would predict virologic success above other variables that are known to be important for virologic success. The investigators analyzed the drug concentrations in 10-20 strands of hair (about 2mg) and using multivariate logistic regression, estimated the effect of hair drug concentrations on a dichotomous measure of virologic success (achieving viral load undetectability or a 1-log drop by 6 months). The results showed that hair concentrations of LPV and RTV were strong and independent predictors of virologic success and may offer a noninvasive tool for determining treatment outcomes. Although at first blush these data look promising, it is unclear how this strategy will improve the outcomes of patients on HIV PIs. There are no prospective studies demonstrating that TDM improves virologic outcomes. There are many retrospective reports showing strong correlations with plasma drug concentrations and virologic outcomes. Dr. Gandhi's data belong in the retrospective correlation category. However, evaluating hair concentrations of drugs has complications that measuring plasma concentrations of drugs have resolved. We are just going to mention a few for the readers to ponder. (A) How can the investigators report concentrations in μg/mL when they are extracting solid matter? A thorough validation of their method needs to be performed. (B) What do we know about the kinetics of PIs contained in the hair? Is there a gradient in the amount of drug along the hair? Is the hair content of drug higher at the hair tip or at the hair shaft? These need to be determined. (C) Are all hairs the same? What about people with thick curly hair vs. thin, straight hair? Certainly the weight of 20 strands of hair is variable based on hair quality even if controlled for length. (D) What happens to hair content of drugs if the hair is chemically treated? (E) Are these data applicable to men? In conclusion the data presented by Dr. Gandhi need confirmation and the analysis of hair drug content requires independent confirmation before we become excited about a new, noninvasive TDM. In addition, the drug content in hair does not reveal about drug-taking behavior, which is important for therapeutic success.
The use of low dose RTV to boost concentrations of other PIs has made this group of drugs more effective in the treatment of HIV infection. The present FDA approved formulation of RTV is a liquid soft gel capsule which requires refrigeration for long term stability. Recently Abbott Labs introduced a meltrex tablet formulation of LPV/r for clinical use. This formulation demonstrated bioequivalence to the soft gel capsule but was stable at higher temperatures and had more predictable bioavailability during various dietary conditions. In addition, the meltrex formulation appears to result in less gastrointestinal upset. In order to be able to use RTV with other PIs in developing countries where refrigeration may not be available, the development of a RTV meltrex tablet formulation is a high priority. Dr. George Hanna from Abbott Labs presented early bioequivalence data in healthy volunteers with three different meltrex formulations of 100 mg RTV (abstract 52LB). Using the FDA bioequivalence criteria, forms A and B of RTV meltrex were found to be bioequivalent to RTV soft gel, but form C had a Cmax that was too high to meet the bioequivalence criteria. Further studies are ongoing examining long-term stability at various temperatures and humidity conditions for these tablets in order to determine which formulation will be the most suitable for registration. These data were exciting because the use of more lipid friendly protease inhibitors like ATV and DRV will soon become a reality in developing countries where owning a refrigerator is not possible.
The final oral presentation was by Dr. Peter Anderson from the University of Colorado Health Sciences Center (abstract 53). We cannot be completely objective about these data since one of us was a co-author (J Gerber) and the other is a close colleague of Dr. Anderson (J Kiser). Briefly, researchers who measure N(t)RTI triphosphate concentrations of drugs in peripheral blood mononuclear cells (PBMCs) do not take into account that PBMCs consist of multiple cell types and that there may be cell type-specific phosphorylation of nucleoside analogues. Anderson et al. previously reported that ZDV triphosphate and 3TC triphosphate concentrations in PBMCs were significantly higher in CD4 depleted HIV-infected subjects as compared to CD4 preserved HIV-infected subjects (Anderson et al., AIDS 17:2159, 2003). Although the more immunosuppressed HIV-infected subjects probably have a greater percentage of activated CD4 and CD8 cells, the possibility that CD4 cells phosphorylate nucleosides less well is also a possibility. Dr. Anderson, therefore, examined the intracellular ZDV and 3TC triphosphate concentrations in blood from HIV-seronegative healthy volunteers who took ZDV 300 mg and 3TC 150 mg BID for at least seven days. The PBMCs were separated, CD4 depleted and enriched using CD4 magnetic immuno-beads. The results showed that CD4 enriched cells phosphorylate ZDV to a significantly lesser extent than either PBMCs or CD4 depleted fractions (3-4 fold difference). 3TC phosphorylation was slightly (but significantly) greater in CD4 depleted cells as well (1.5-fold higher). In these healthy volunteers, the percent of activated cells in the PBMCs was small (<2%). These data are important because for the first time they demonstrate that nucleoside triphosphate concentrations are not uniformly distributed in the various cells in the PBMCs. Similar data needs to be generated in HIV-infected subjects to evaluate how increased cell activation affects the intracellular triphosphate concentrations. Nonetheless these findings are not enough to explain Dr. Anderson's previous data showing that CD4 depleted HIV-infected subjects show much higher ZDV phosphorylation. More work needs to be generated in this field in order to get a better understanding of the efficacy and toxicity of the N(t)RTI's.
The rest of the overview will include selected abstracts that we found interesting and we think have clinical relevance.
Drug:Drug Interactions
Rifampin is a potent inducer of many enzymes and transporters, and thus its concomitant use with ARV drugs can compromise ARV drug efficacy. Several QD ATV/r dosing strategies were previously tried in an effort to overcome rifampin induction, however these efforts were unsuccessful (Burger et al., AAC 50:3336, 2006). ACTG 5213 evaluated BID unboosted ATV in combination with rifampin in healthy volunteers (abstract 575, Acosta et al.). The logic behind this study was that ATV demonstrates saturation pharmacokinetics, thus a higher than usual daily dose given BID may withstand rifampin's strong induction of metabolism. It would be important to find a HIV PI that could be used in developing countries where tuberculosis (TB) is the most common opportunistic infection in HIV-infected people and adequate therapy for both diseases is vital. Neither the 300 nor 400 mg BID doses maintained adequate ATV exposures with rifampin. Fortunately, the volunteers tolerated the combination well, unlike the previous studies with SQV/r and LPV/r with rifampin in which volunteers developed significant hepatotoxicity. PK studies of BID ATV/r with rifampin are being developed by this group.
Developing countries frequently have no choice but to use rifampin and NVP concomitantly in HIV/TB coinfected patients because rifampin is the only rifamycin at their disposal and NVP is part of a generic fixed dose combination product. Therefore determining the appropriate dose of NVP in the setting of rifampin in these countries is imperative. Abstract 576 reported by Avihingsanon et al. studied 30 HIV/TB-coinfected Thai patients on rifampin who were randomized to receive NVP 400 mg daily or 600 mg daily. The NVP median AUC, Cmax, and Cmin were significantly higher in those on 600 mg QD, but there was no difference in CD4 cell count (76 vs. 88) or proportion with HIV-RNA less than 50 copies/mL (53 vs 43%) between dosing groups. Of note, 10 patients (5 in each group) had discontinued study by week 12. Four patients (all on 600 mg QD) developed severe rash and fever. The optimal dose of NVP (if there is one) in the setting of rifampin is still unknown. In choosing a dose of NVP, safety and efficacy are the main concerns and at this point neither have been definitively established.
Rosuvastatin is a potent agent for reducing LDL cholesterol. 90% of a rosuvastatin dose is excreted unchanged through biliary secretion, thus an interaction with PIs (which are potent inhibitors of CYP3A4/5) was not expected. Additionally, previous interaction studies with other CYP450 inhibitors have shown no effect on rosuvastatin PK (Cooper et al. Eur J Clin Pharmacol 58:527, 2002, Cooper et al. Br J Clin Pharmacol 55:94, 2003). Rosuvastatin is the most potent statin in clinical use and it would be a welcome addition to the armamentarium of lipid lowering agents for HIV-infected subjects on HAART. A study presented by Kiser et al (abstract 564) showed that in the presence of LPV/r, rosuvastatin Cmax and AUC were increased 4.7- and 2.1-fold. Despite the increase in exposure, the lipid lowering effect of rosuvastatin was attenuated in the presence of LPV/r. The investigators speculate that the mechanism for this interaction is via LPV or RTV (or both drugs) inhibition of organic anion transporting polypeptide 1B1 (OATP1B1). OATP1B1 is a basolaterally localized hepatic transporter for which rosuvastatin is a high affinity substrate. The implications of this purported mechanism are that rosuvastatin concentrations will be elevated, but that the lipid lowering efficacy of rosuvastatin will be reduced because rosuvastatin is not penetrating the liver - where it would exert its LDL-lowering effects. Additional studies are needed to determine the exact mechanism for this interaction. In the meantime the concomitant use of rosuvastatin and LPV/r requires careful surveillance.
A drug-drug interaction study of artesunate plus amodiaquine (antimalarial agents) with EFV in healthy volunteers was prematurely discontinued after two patients developed significant liver function test (LFT) elevations (abstract 577, German et al.). The LFT elevations were likely secondary to significantly increased amodiaquine levels; probably due to EFV's inhibition of CYP2C8, the enzyme responsible for amodiaquine oxidation. The investigators report that there is an in vitro study in press confirming this mechanism for the interaction. An artemether/lumefantrine interaction study with EFV is actively recruiting, and to date they have not seen LFT elevations as with artesunate plus amodiaquine and EFV. It is exciting to finally see drug-drug interaction studies with antimalarial and ARV agents since malaria and HIV co-infection are commonly observed in developing countries.
Diltiazem, given as 240 mg long acting tablets QD, and itraconazole given as 200 mg BID, AUCs were reduced 76 and 39% respectively by EFV in two healthy volunteer studies. Thus, a higher dose of diltiazem and a higher dose of itraconazole, or an alternative antifungal, may be required when these drugs are co-administered with EFV (abstract 561, Kaul et al.).
Minocycline and valproic acid are being investigated as adjunctive therapies for HIV-associated cognitive impairment. Valproic acid is also utilized as a mood stabilizer in HIV-infected subjects with bipolar disease. In a study of 12 HIV infected adults on a stable ARV regimen which included ATV/r 300/100 mg QD, minocycline reduced ATV AUC and Cmin by 33 and 50%, respectively. When valproic acid was added to minocycline, the PK effect on ATV was maintained. The potential mechanism by which minocycline lowers ATV levels is unknown and from this study an independent effect of valproic acid on ATV AUC cannot be ascertained (abstract 567, DiCenzo et al.).
Both H2 blockers and TDF lower the concentrations of ATV (Agarwala et al. 6th PK Workshop, abstract 11, Taburet et al., AAC 48:2091, 2004). In anticipation of an additive interaction upon simultaneous administration of ATV/RTV/TDF and famotidine, different dosing regimens of famotidine were explored in healthy volunteers (abstract 568, Agarwala et al.). The administration of 20 mg of famotidine BID with either simultaneous morning administration with ATV/RTV/TDF or temporal separation did not significantly alter ATV AUC or Cmax, Cmin was reduced approximately 20% for both strategies. Morning administration of ATV/RTV/TDF and evening administration of famotidine 40 mg QD resulted in an 11, 12, and 23% decrease in ATV Cmax, AUC, and Cmin. Famotidine 40 mg BID taken 2 hours after ATV/RTV/TDF in the morning resulted in a 26, 21, and 28% reduction in ATV Cmax, AUC and Cmin. From these data it appears that the use of TDF and famotidine does not decrease ATV concentrations anymore than use of an H2 blocker alone. Temporal separation and use of the lowest possible dose of the H2 blocker minimize the reduction in ATV exposures.
Pharmacokinetic Issues in Hepatitis Coinfection
Hepatic impairment can alter the PK of ARV drugs metabolized by the liver. These data were recently reviewed by one of us (Wyles and Gerber, CID 40:174, 2005). Abstract 946 by Breilh et al. presented Cmax and Cmins for RTV-boosted LPV, ATV, and fAPV, and EFV, and NVP in 113 HIV/Hepatitis C Virus (HCV) coinfected patients. The Cmax and Cmin were compared between patients with a Fibroscan of less than or greater than 12 kPa (which is considered the cut off value for cirrhosis). The Cmax for LPV, ATV, fAPV were obtained at 3, 3, and 2 hours post-observed dose respectively. Neither Cmax nor Cmin differed by liver stiffness score for the RTV-boosted PIs. This is surprising since ATV, LPV/r, and APV AUCs have been shown to be increased in patients with liver dysfunction in previous studies (Reyataz prescribing information, Peng et al. J Clin Pharmacol 46:265, 2006, Lexiva prescribing information). The inability to find a difference in this study in Cmax and Cmin values in patients with and without cirrhosis is likely a result of insufficient power to detect a difference due to the large interpatient variability in Cmin and Cmax concentrations and the small number of subjects on each drug. However Cmin's for EFV and NVP were significantly higher in those patients with Fibroscan scores > 12 kPa. A more intensive PK study in subjects with various stages of liver disease would likely provide a better understanding of the true impact of liver dysfunction on these ARV drugs. (note from Jules Levin: perhaps Fibroscan evaluation of cirrhosis was off).
Despite many promising investigational therapies against HCV, it is likely that these agents will be used in combination with pegylated interferon and ribavirin in order to limit the development of resistance. Thus, we still need to learn how to optimally dose ribavirin and interferon in order to maximize efficacy and minimize toxicity. Several posters were presented at this year's CROI on TDM of ribavirin and predictors of pegylated interferon/ribavirin treatment success. Ribavirin plasma and erythrocytes Cmins were correlated with virologic response and hemoglobin declines in 22 patients (abstract 893, Dominguez et al.). In a separate study in 42 patients, a ribavirin Cmin of ≥1600 ng/mL was associated with early and sustained virologic response, while a ribavirin Cmin > 2300 ng/mL seemed to predict toxicity (abstract 903, Marucco et al.)
ABC-based ARV therapy was identified as a risk factor for failure to achieve an early HCV virologic response in HIV/HCV coinfected patients on pegylated interferon alpha 2b plus ribavirin 800 mg daily (abstract 897, Bani-Sadr, et al.). This is an interesting finding that seems biologically plausible based on the potential pharmacologic antagonism of these drugs, but obviously requires further study (note from Jules Levin: I was taking HAART containing abacavir & tenofovir while on Peg/RBV & achieved undetectable HCV RNA within 5 weeks & SVR, so I'm not sure I believe this abstract. It did not seem like a well done study). Alternatively, TDF or d4T plus 3TC were associated with a higher rate of sustained virologic response to pegylated interferon plus ribavirin (abstract 898, Mira et al.). Two studies found ZDV in combination with ribavirin to be significant predictors of severe toxicity (abstract 902, Nunez et al. and abstract 904, Mira et al.).
Antiretroviral Pharmacokinetics in Pregnant Women
LPV/r and NFV are the preferred PIs in the United States Department of Health and Human Services Guidelines (DHHS) for the treatment of pregnant HIV-infected women. LPV protein binding in HIV-infected pregnant women was described in 28 women participating in Pediatric AIDS Clinical Trials Group study (PACTG) 1026 (abstract 787, Aweeka et al.). Most of the HIV PIs are highly bound to alpha-1 acid glycoprotein (AGP). In this study, AGP concentration was lower and LPV fraction unbound was higher during the third trimester of pregnancy compared to post-partum. The reported 17% relative increase in LPV fraction unbound during pregnancy is much smaller than the reported reduction in total LPV concentrations (Stek et al. AIDS 20:1931, 2006). Therefore, reduced protein binding does not account for most of the reduction in LPV concentrations during pregnancy.
NFV 250-mg tablets dosed at 1250 mg BID previously demonstrated acceptable PK in HIV-infected pregnant women. However, data presented from PACTG 1026 suggested NFV concentrations were lower during the third trimester vs. post partum in women receiving the 625-mg NFV tablet formulation at 1250 mg BID (abstract 740, Read et al.). The median NFV AUC, Cmax, and Cmin in 25 women during the third trimester were 18.7 μg*hr/mL, 3.1 μg/mL, and 0.5 μg/mL. The 50th percentile for NFV AUC and Cmin in non-pregnant patients are 29 μg*hr/mL and 0.7 μg/mL, respectively. Despite the lower concentrations of NFV, 21 of 22 women during the third trimester had viral loads of < 400 copies/mL. Additional studies are needed to determine the long term implications of these subtherapeutic NFV concentrations during pregnancy, and if a higher dose of NFV may be required. In addition, the effect of pregnancy on the protein binding of NFV may yield useful results since NFV is very highly protein bound in plasma.
SQV-HGC (as Invirase) is an alternative PI in the DHHS guidelines. One study found the concentrations of SQV/r 1000/100 mg BID to have comparable PK in HIV-infected women during their third trimester to historic data (abstract 741, Burger et al.). However, abstract 753 (Hanlon et al.) found that 13 of 42 women who initiated an ARV regimen including SQV/r 1000/100 mg BID experienced elevated transaminases (grade 1-3) shortly after initiating their regimens.
Presently there are very few data on the PK, safety, and efficacy of TDF in HIV-infected women. PACTG 1026 investigated TFV concentrations in 19 HIV-infected pregnant women. TFV AUC, Cmax, and Cmin were decreased 10-20% in the third trimester of pregnancy compared to post-partum, but 88% of women in the third trimester had undetectable viral loads (Abstract 738b, Burchett et al.). Both this study and an Italian study (abstract 738a, Bonora et al.) also showed that TFV had an efficient transplacental passage with the ratio of maternal to umbilical cord TFV concentrations at delivery being very close to 1. So far there are no reports of safety concerns with TDF to the fetus but close surveillance needs to be maintained.
Data are also limited on the use of ATV in pregnancy, and there are concerns of neonatal hyperbilirubinemia with this drug. Two studies at CROI reported on the use of ATV in pregnancy (abstract 742, Ripamonti et al. and abstract 750, Natha et al.). When ATV/r 300/100 mg QD was used in nine HIV-infected pregnant women, ATV AUC, Cmax, and Cmin were only approximately 6% lower in third trimester vs. postpartum. In a separate study, 26 infants were born to HIV-infected women taking ATV in combination with other ARV agents (abstract 750). The median bilirubin level in the infants was 4.2 mg/dL, none required phototherapy and no birth defects were detected. Although these preliminary data are encouraging on the overall safety of ATV in pregnancy, additional PK and safety data are still necessary.
Pharmacokinetic Issues in Children
It is encouraging to see more PK data on ARV drugs in children. Essentially all the pediatric HIV infections are in developing countries and the future generations in those countries depend on the correct use and availability of ARVs for treatment and prevention of HIV infection.
Triomune (Cipla) fixed-dose adult antiretroviral combination preparation containing 3TC, d4T, and NVP was previously shown to provide subtherapeutic NVP concentrations to children receiving the combination. Pedimune Baby and Junior (Cipla) are generic fixed dose combination tablets developed for pediatric use each containing stavudine/lamivudine/nevirapine 6/30/50mg and 12/60/100mg, respectively. 35 children were dosed according to weight to evaluate the PK of d4T, 3TC, NVP in the fixed dose pediatric combination preparations (abstract 580, Kabamba et al.). 3TC and d4T concentrations were similar to historic values in adults, however NVP concentrations were slightly higher than some historical adult values. Three subjects had a grade 2 rash, 3 had grade 3 elevations in liver function tests, and 1 had a grade 1 rash that progressed to a grade 3. The investigators did not correlate NVP PK with adverse events. This study illustrates well the problems we face in trying to use fixed dose combination tablets in children. To use these fixed dose combination tablets, we will have to find the balance between achieving adequate concentrations and minimizing toxicities for all the individual components, which is certainly a challenge.
Ren et al. presented an oral abstract on LPV/r and EFV concentrations in children in the presence and absence of rifampin in South African children (abstract 77). Three EFV concentrations were obtained between 12 and 24 hours post dose for HIV/TB coinfected children on standard dose EFV both with and without rifampin coadministration and in an additional control group which included HIV-infected children without TB on standard dose EFV. Cmins were estimated from the half-life calculated using the three time points. The TB coinfected and control groups were similar in age. EFV Cmins were not statistically different between the three groups, however 50% of the children had EFV Cmins of less than 1 μg/mL which may eventually result in the development of NNRTI resistance. In the study of LPV/r with and without rifampin, 13 TB coinfected children received LPV/r in a 1:1 dosing ratio based on a previous study (la Porte et al. AAC 48:1553, 2004) in combination with rifampin. The control group consisted of HIV-infected children without TB on standard dose LPV/r. All children underwent intensive 12 hour PK studies for LPV/r plasma concentrations. LPV AUC and Cmax were significantly reduced by 25 and 16%, respectively in children on rifampin despite the LPV/r dose adjustment. The Cmin was not statistically different, but there was large interpatient variability. The authors concluded that since Cmin was not different and because all 28 subjects had troughs > 1 μg/mL, that this LPV/r dosing strategy appeared to overcome the rifampin induction. We think that this is a premature recommendation because there is such large interpatient variability in the plasma LPV concentration reductions secondary to rifampin induction and no tolerability or toxicity data were reported. Since la Porte et al. found significant hepatotoxicity in seronegative adult volunteers when rifampin was combined with higher doses of LPV/r, careful toxicity surveillance is critical with these drug combinations.
It appears from data presented by Rutstein et al. that we are close to determining the optimal dose of ATV and ATV/r in children (abstract 715). PACTG 1020 is a Phase I/II trial of both boosted and unboosted ATV in children 91 days to 21 years in the United States and South Africa designed to determine the safety, PK, and optimal dosage of ATV powder and capsules. ATV, like other PIs, appears to require a higher dose on a mg/m2 basis in children compared to adults, especially when it is not RTV-boosted. Doses of 520 mg/m2 and 620 mg/m2 of unboosted ATV capsules met the protocol-defined PK criteria in children 2 to 13 years, and 13 to 21 years, respectively. Unboosted ATV powder did not achieve the protocol-defined PK targets, thus RTV boosting is required when administering the powder. At week 24, 79% of treatment-na•ve and 43% of treatment-experienced children had viral loads < 400 copies/mL. Interestingly, trough concentrations, but not AUCs or Cmaxs were significantly lower in South African children than American children. This could be the result of different diets or a higher gastric pH in South Africans, but this finding requires further examination.
Eight HIV-infected infants between 2 and 6 weeks of age received LPV/r 300/75 mg/m2 twice daily plus two NRTI (abstract 716, Pinto et al.). Intensive 12 hour PK data were obtained after 2 weeks on LPV/r for seven infants. At a median of 7.2 weeks of age, the median LPV AUC, Cmax, and Cmin were 33 μg*hr/mL, 3.8 μg/mL, and 2.6 μg/mL. These values are significantly lower than historic values in adults. Despite this, only one infant reached a predefined virologic failure at week 24 due to medication non-adherence. Additional longitudinal data are necessary to understand the long-term implications of the lower LPV exposure in this age group.
Miscellaneous Topics
This final section is being named miscellaneous topics for the sake of brevity. In this section we will just give punch lines on a few abstracts that we thought were worthy of a mention.
1. Bertz et al. (Abstract 565) examined the PK/pharmacodynamic relationship with ATV from the BMS 089 study that compared the efficacy and safety of ATV/r (300/100 mg) QD to ATV 400 mg QD combined with 3TC and d4T XR in treatment-na•ve subjects. The efficacy data has been previously presented and showed that ATV 400 mg QD was non-inferior to ATV/r. What was remarkable in this study was that the vast majority of subjects achieved HIV-RNA <400 copies/mL by week 48. Although ATV Cmin was associated with the probability of achieving HIV-RNA <400 c/mL and <50 c/mL, even with the ATV Cmin in the lowest quartile (9.2-118 ng/mL) the percent of patients with HIV-RNA <50 c/mL was 74% and in the highest quartile (764-2488 ng/mL) the percent with HIV-RNA <50 c/mL was 89%. Total bilirubin and probability of jaundice were positively correlated with Cmin, but lipid changes to the regimen had a weak correlation to ATV Cmin. Although these data were interesting, it would have been of interest to know the overall adherence and selective drug adherence in subjects with very high ATV Cmin but still detectable HIV-RNA.
2. Previous studies have suggested sex-related difference in ARV PK. Adult AIDS Clinical Trials group study 5223 evaluated the potential for sex-related differences in LPV and RTV PK in 78 HIV infected subjects receiving LPV/r soft gel capsules as part of their ARV therapy (abstract 786, Umeh et al.). LPV PK did not significantly differ between men and women, however RTV AUC and Cmax were 25 and 40% higher in women compared to men. Sex differences in LPV PK using the meltrex formulation are presently being investigated.
3. Single nucleotide polymorphisms (SNPs) in CYP2B6 have been well described in American and European populations, however little is known about the frequency of these SNPs in other populations (Ribaudo et al. CID 42:401, 2006, Haas et al. JID 192:1931, 2005). Two studies (abstract 553, Jamshidi et al. and abstract 554, Ramachandran et al.) evaluated CYP2B6 allele frequencies in African and Indian subjects. CYP2B6 is the main enzyme involved with EFV and NVP metabolism, drugs commonly used in developing countries. Indian CYP2B6 516T homozygous variant carriers had significantly higher EFV and NVP concentrations than wild-type or heterozygotes. What was surprising was that the frequency of CYP2B6 516T variant carriers was much higher in Indian subjects than the frequency previously described in Caucasian and African-American subjects (40 vs. 3 and 20%, respectively). Additionally, high rates of CYP2B6 SNP variant carriers were identified in Ugandan and Zimbabwean patients. These data have implications for EFV and NVP dosing and potential adverse effects in various patient populations.
4. Etravirine (TMC125), a twice daily NNRTI with activity against wild-type and K103N resistant virus, is available as expanded access. Pharmacokinetic/pharmacodynamic relationships with etravirine from the TMC125-C223 study in 199 HIV-infected patients with NNRTI resistance and at least 2 primary PI mutations were explored(abstract 560, Kakuda et al.). Pharmacokinetic data were available for 150 subjects. No significant association between TMC125 pharmacokinetic parameters and age, weight, gender, race, or viral hepatitis status was observed. However, use of lopinavir/ritonavir, tenofovir disoproxil fumarate, and the 400 mg twice daily TMC125 dose were associated with a statistically significant lower TMC125 AUC and Cmin. Tenofovir disoproxil fumarate and lopinavir/ritonavir use were associated with a 43% and 24% decrease in TMC125 AUC. TMC125 AUC and Ctrough and de-novo enfuvirtide use were identified as significant factors for viral load response. No relationship was found between TMC125 AUC and clinical or laboratory adverse effects. This study used the older formulation of TMC125 at doses of 400 mg twice daily and 800 mg twice daily. Phase III studies are using a newer formulation of TMC125 that has much improved bioavailability at a dose of 200 mg twice daily.
CONCLUSIONS
We tried to cover many pharmacology topics from 2007 CROI and we neglected some because of space issues. We feel that we covered the most important and controversial topics. We tried to analyze the data objectively but no doubt personal bias was introduced during data interpretation. Drug-drug interactions once again took center stage at this meeting. Trying to determine how to use rifampin with ARV drugs remains a high priority since in developing countries TB is the most common opportunistic infection seen with HIV infection. The interaction between rosuvastatin and LPV/r was unexpected - once again pointing out that understanding the role of transporters in drug elimination remains important. There is more to drug disposal than metabolism. And finally clinical pharmacology in the pediatric population received some much-deserved attention at this meeting. This is critical since the future of many developing countries depends on providing safe and effective therapy to both infants and children born to HIV-infected mothers. An overall understanding of ARV drug pharmacology is the only way to optimize efficacy and minimize toxicity of drugs in order to improve clinical outcomes.
|
| |
|
|
|
|
|
