Clinical Pharmacology at CROI 2015
Courtney V. Fletcher, Pharm.D.
Dean and Professor
Anthony T. Podany, Pharm.D.
College of Pharmacy
University of Nebraska Medical Center
986000 Nebraska Medical Centerv
Omaha, NE 68198
The 2015 (22nd) Conference on Retroviruses and Opportunistic Infections (CROI) was held in Seattle, WA, from February 23 through February 26, 2015. CROI continues to be the premier HIV-focused scientific meeting. In this report we will highlight abstracts focused on pharmacologic issues that are of broad interest or might benefit from some expert clarification. Abstracts will be discussed in the broad categories of: (i) the therapy of HIV infection, (ii) PrEP, and (iii) drug-drug interactions.
%CV, percent coefficient of variation
ACTG, adult AIDS clinical trials group
ARV, antiretroviral drug
ART, antiretroviral drug therapy
AUC, area under the concentration-time curve
BID, twice daily
C12, drug concentration at 12 hours post dose
Cmax, maximum drug concentration
Cmin, minimum drug concentration
CNS, central nervous system
CSF, cerebrospinal fluid
CVF, Cervicovaginal fluid
Ctrough, concentration immediately before the next dose
CYP, cytochrome P450 drug metabolizing enzymes
DBS, dried blood spot
DHHS, Department of Health and Human Services
DSMB, data safety monitoring board
GMR, geometric means ratio
HAND, HIV-associated neurocognitive disorders
HDAc, histone deacetylase
IC50, concentration of drug required to inhibit viral replication in vitro by 50%
IC90, concentration of drug required to inhibit viral replication in vitro by 90%
IQ, inhibitory quotient
IVR, intra-vaginal ring
MPA, medroxyprogesterone acetate
NRTI, nucleoside reverse transcriptase inhibitor
NNRTI, non-nucleoside reverse transcriptase inhibitor
PACTG, pediatric AIDS clinical trials group
PBMCs, peripheral blood mononuclear cells
PEG-IFN, pegylated Interferon
PI, inhibitor of HIV protease
PrEP, pre-exposure prophylaxis
QD, once daily
r or RTV, ritonavir
TAF, tenofovir alafenamide fumarate
TDF, tenofovir disoproxil fumarate
TDM, therapeutic drug monitoring
I. The Pharmacotherapy of HIV Infection
CROI has long been the premier venue for presentation of clinical trials that advance the treatment of HIV infection. These trials are covered in depth elsewhere on the NATAP http://natap.org website. Here, we will give some bullet points and provide some pharmacologic commentary.
Tenofovir vs. Tenofovir (or TAF vs. TDF, Abstracts 113LB and 143LB).
1733 HIV-infected persons were enrolled in 2 clinical trials and randomized to receive TAF or TDF in addition to EVG/COBI/FTC. At the 48 wk primary endpoint, TAF was non-inferior to TDF, with 92% and 90% of participants, respectively, having HIV-RNA in plasma < 50 copies/mL. There were no reports of proximal renal tubulopathy in either arm. Adverse effects on serum creatinine and lumbar spine and total hip bone mineral density (BMD) were less in TAF vs TDF recipients. Increases in fasting lipids were greater in TAF recipients. The mean plasma TFV AUC was 297 vs 3410 ng*h/mL in TAF vs TDF recipients.
Commentary. In these Phase 3 clinical trials, TAF was non-inferior to TDF, and the renal and BMD safety profile was better. This appears to arise because TAF is able to achieve higher intracellular concentrations of TFV-diphosphate (the pharmacologically-active moiety) at lower plasma TFV concentrations. We don't see, however, how lower plasma concentrations explains the greater increase in fasting lipids seen with TAF. What is important to emphasize is that while TAF has a better renal and BMD safety profile, it still had adverse effects, e.g. a 6.6 ml/min decrease in creatinine clearance vs 11.2 for TDF over 48 wks. Here is a useful reminder from the original FDA-approved package insert for TDF from October 26, 2001: "Although tenofovir-associated renal toxicity has not been observed in pooled clinical studies for up to one year, long term renal effects are unknown." Those first reports of tenofovir-associated renal toxicity began to appear about one year after approval. TAF appears to be an important advance: equally efficacious and less toxicity compared with TDF. But, these are still the early days of TAF and the full safety profile will not be known until it is used in a larger number of persons for a longer period of time.
CROI: Tenofovir Alafenamide (TAF) in a Single-Tablet Regimen in Initial HIV-1 Therapy - (02/26/15)
CROI: Safety of Tenofovir Alafenamide in Renal Impairment - (02/27/15)
CROI: Renal and Bone Safety of Tenofovir Alafenamide vs Tenofovir Disoproxil Fumarate - (02/27/15)
CROI: Safety of Tenofovir Alafenamide in Renal Impairment - (02/27/15)
Antiretrovirals and Kidney Disease (Abstracts 142 and 792, 793, 794 and 795).
The Data Collection of Adverse Events of Anti-HIV Drugs (D:A:D) international collaboration reported on the development of chronic kidney disease (CKD) in 23,560 HIV-infected persons with normal renal function at baseline followed for a median of 6.3 years (#142). The overall incidence of CKD was 0.9%.
Increasing exposure to TDF, ATV/r and LPV/r was associated with a significant increase in CKD. The relative risk of CKD was increased 97%, 320% and 140% for TDF, ATV/r and LPV/r, respectively, after 6 years exposure. Only discontinuation of TDF use was associated with a decrease in the incidence of CKD although it remained higher for 2 years compared with those who never started TDF. See a complementary abstract (#794, poster) that describes a relationship among TDF and ATV/r with acute kidney tubular injury.
CROI: Exposure to antiretrovirals and development of chronic kidney disease - D:A:D Study Group
The Women's Interagency HIV Study (WIHS) performed an intensive PK study in 105 HIV-infected women receiving TDF (#792, poster). Over 7 years of followup, higher TFV concentrations were associated with a decline in estimated glomerular filtration rate (eGFR). Those women who had the highest TFV plasma concentrations had significantly lower eGFR at the time of the intensive PK study (baseline) and at year 7.
A strong genetic susceptibility to CKD has been observed in the sub-Saharan African population. Investigators described the effect of 2nd line ART on renal function in 443 persons from 3 African countries who had failed first line therapy (#793, poster). These participants were randomized to one of 3 regimens: (A), TDF+FTC+LPV/r; (B), ABC+ddI+LPV/r; or (C), TDF+FTC+DRV/r and were followed for 18 months. All 3 groups had a marked decline in eGFR over the first month; the largest decrease in eGFR was in the TDF+FTC+LPV/r recipients. At month 18, mean eGFR in the non-TDF containing regimen of ABC+ddI+LPV/r had recovered to baseline eGFR, and was significantly greater than those in the 2 TDF-containing arms. The incidence of CKD was 5%, 5% and 4% person-year in arms A, B and C, respectively.
242 patients with moderately reduced renal function (eGFR between 30 to 69 mls/min) had their treatment switched from a TDF-containing (65% of participants) or a non-TDF-containing regimen to TAF-containing regimen with EVG+cobi+FTC (#795). Baseline median eGFR was 55.6 mls/min and was not significantly different at week 24 where eGFR was -0.4 ml/min lower. Actual GFR was measured in a subset of 32 participants and no difference was found between baseline and week 24 (57 vs 59 mls/min, respectively). The prevalence of clinically significant proteinuria and albuminuria decreased from baseline to week 24 as did markers tubular function markers retinol binding protein and beta-2 microglobulin.
[CROI: Safety of Tenofovir Alafenamide in Renal Impairment]
Commentary. Collectively, these abstracts tell us that certain ARVs are associated with drug-induced kidney disease. The WIHS abstract indicates (as has other publications) a link between increased plasma concentrations of TDF and reduced kidney function. This underscores the need for proper TDF dose adjustment based on renal function. The incidence of CKD is low, but certain populations may have an increased risk and consideration for using more renal-friendly ARVs is appropriate. The switch to TAF in patients who already had existing renal impairment can be taken as support that a reduction in TFV plasma concentrations improves renal safety. But as mentioned above, we are still in the early days of learning about the safety profile of TAF.
Safety of Strategies to Prevent Perinatal HIV Transmission (PROMISE, Abstract #31LB, webcast).
PROMISE is an ongoing trial to assess strategies to prevent mother-to-child transmission (PMTCT). At CROI 2015, data on the efficacy and safety of ARVs during the pregnancy through day 14 postpartum period were presented. In this period, women were randomized to 1 of 3 regimens: (A) ZDV + single dose NVP and FTC/TDF to cover the tail; (B) ZDV/3TC + LPV/r; or (C) TDF/FTC + LPV/r. 3529 pregnant women enrolled in the study and data on 3234 live births were reported. The MTCT was significantly lower with the 2 triple therapy arms (B and C) compared with ZDV + sdNVP: 0.56% rate of infant infections for B and C vs 1.8% for A. Moderate adverse pregnancy outcomes (low birth weight and preterm delivery) were higher with the ZDV/3TC + LPV/r than Arm A (Arm C could not be statistically evaluated because of study design issues). Rates of severe pregnancy outcomes (very low birth weight and very preterm delivery) were lower with ZDV/3TC + LPV/r than with TDF/FTC + LPV/r. There was also a significantly lower risk of infant death by age 14 days for ZDV/3TC + LPV/r (0.6%) than with TDF/FTC + LPV/r (4.4%).
Commentary. The WHO has recommended triple therapy ARV regimens in pregnancy since 2013. These results provide strong support for that recommendation; there was approximately a 3x lower rate of infant HIV infection with triple therapy. In general, these ARVs are remarkably safe in pregnant women and their infants - remember the mechanism of action of NRTIs is termination of DNA synthesis and the safety of these agents is dependent on their ability (that does vary by NRTI) to selectively inhibit replication of HIV and not inhibit human mitochondrial DNA polymerase, which causes of potentially serious toxicity (neutropenia, lactic acidosis, pancreatitis, etc). The differences observed in severe pregnancy outcomes and infant death favoring ZDV/3TC + LPV/r over TDF/FTC + LPV/r were surprising, especially given the well-documented superior safety of TDF/FTC over ZDV/3TC in adult clinical trials. That said, these results and adverse safety signals cannot be discounted; they require additional investigation and should be considered when selecting a triple therapy regimen for PMTCT.
Ia. HIV Associated Neurocognitive Disorders (HAND) and Antiretroviral CSF Pharmacology
Abstract 84 [webcast] investigated the correlation between efavirenz (EFV) concentrations its major metabolite, 8-hydroxyefavirenz (8-OH-EFV) and neuropsychological performance (NP) in older HIV-infected patients. Better neurocognitive functioning was associated with higher EFV metabolite (8-OH-EFV) concentrations, but not plasma EFV concentrations. Additional studies of EFV and it's metabolite concentrations may improve our understanding of antiviral effect and risk for neurotoxicity.
Abstract 57 [webcast] examined NP in patients from sub-Saharan Africa who failed first line ART and their responses to second line ART. 1036 patients had baseline (at time of switch to 2nd line therapy) NP test available while 915 patients had week 96 NP data available. Patients who failed 1st line ART had poorer NP function, likely related to uncontrolled HIV infection, while NP greatly improved upon second line therapy. These results reinforce the association between improved NP in individuals with controlled viremia.
Abstract 56 [see link below] reported results of the first randomized control trial (RCT) designed to compare efficacy of ART regimens in preventing neurocognitive (NC) decline. The authors enrolled 250 HIV+, ART naïve adults in Beijing with baseline CD4+ counts below 350/mm3. All subjects were considered NC normal at baseline. Participants were randomized to either zidovudine-lamivudine-nevirapine (ZLN) or tenofovir-lamivudine-efavirenz (TLE) and extensively followed for 96 weeks with a battery of NC tests. The TLE regimen was associated with a greater risk of incident NC impairment over the 96 weeks than the ZLN group (p=0.009, ITT, p=0.037 AT). There was far greater discontinuation of study drug due to adverse events in the ZLN compared with the TLE group (38% vs. 2%). A nested case-control study within the trial (Abstract 444) demonstrated NC impairment was associated with TDF and EFV drug concentrations. Participants described as decliners in NC impairment had lower CSF-to-plasma ratios of TDF while trending toward higher with EFV. These data may point toward suboptimal TFV concentrations in the CSF, while the higher EFV concentrations may be having a neurotoxic effect. Collectively, the results from these studies indicate that not all regimens are equal in their ability to prevent NC decline, and that further research is necessary to find the optimal regimen that will be effective, tolerable, and have the best NC profile.
CROI: Neurocognitive Decline is Associated with Antiretroviral Concentrations in Cerebral Spinal Fluid and Blood -
Abstract 441 investigated whether adding maraviroc to patients on stable ART with HAND would have an impact on neurocognitive performance. In this small study, 14 HIV+ males with recent progression to HAND were randomized 1:1 to either continue their existing ART or intensify with maraviroc. Although not powered to detect statistically significant effects, the authors did find a clinically relevant effect of improved global neurocognitive performance by intensifying with maraviroc. This effect will need to be evaluated in larger studies.
CROI: Maraviroc-Enhanced cART Improves Cognition in Virally-Suppressed HAND: A Pilot Study - (03/27/15)
Abstract 436 presented ARV concentrations in brain tissue collected from 9 HIV infected adults within 6 months of their death. Interestingly, human brain concentrations of ATV, EFV, FTC and 3TC were similar to literature reported CSF values. However, TFV brain concentrations were higher than reported CSF values while LPV concentrations were higher only in the white matter portion of the brain as compared with CSF. CNS disposition data such as these will help inform future studies targeting both antiviral efficacy and associated CNS toxicity.
CROI: Antiretroviral Concentrations in Brain Tissue Are Similar to or Exceed Those in CSF - (03/23/15)
Commentary. We were happy to see an increase in work on HAND and CNS pharmacology at CROI 2015. One particular clinical need is how to design a CNS-targeted antiretroviral regimen that more fully protects the CNS from the effects of HIV-infection.
Ib. ART, Pharmacology and Resistance
Abstract 117 was a retrospective analysis of patients with low-level viremia (50-999 HIV RNA copies/mL) while on ART. The authors consented 328 patients whom had untimed drug levels (UDL) and genotype data corresponding with the time of first documented low-level viremia. Drug concentrations were categorized as either 'optimal' or 'subtherapeutic' according to recommended trough levels by the DHHS guidelines. A genotype susceptibility score depicted the number of active drugs in the patient's regimen corresponding with their genotype (ie GSS 3 has 3 active drugs). A single UDL was better able to predict subsequent virologic failure than the genotype susceptibility scores (Adjusted Hazard Ratio 2.53 vs 1.55, UDL vs GSS<3). These data reinforce the predictive value of subtherapeutic drug concentrations in virologic failure. Coupled with the demonstrated role of measuring ARV concentrations to identify poor adherence as shown in PrEP trials, is it time to revisit TDM as a tool to identify persons who have low concentrations and are therefore at risk, for example, of acquisition of HIV infection or virologic failure?
CROI: Untimed Drug Levels & Resistance in Patients Experiencing Low-Level Viremia -
II. Pre-Exposure Prophylaxis (PrEP)
There was much presented at CROI 2015 on the topic of pre-exposure prophylaxis (PrEP). Raphael Landovitz's fantastic plenary talk kicked things off by reviewing how far the field has come to date, while new PrEP data was rolled out in the oral and poster abstracts. You can access Landovitz's talk through the
CROI webcasts: http://www.croiwebcasts.org/console/player/25535?mediaType=audio&
Tuesday morning's oral sessions delivered a host of PrEP clinical trail data mostly positive. The few disappointing results were similar to what we have seen in failed PrEP studies in the past where poor adherence to study drug caused a lack of efficacy. First, the disappointing results. The FACTS 001 Trial of pericoital tenofovir 1% gel for HIV prevention in women (Abstract 26LB webcast concluded that 1% TFV gel was not effective at preventing HIV acquisition. The authors found an association between adherence and effectiveness, similar to gel products tested previously. In our opinion, future development of topical microbicides must demonstrate a high degree of patient acceptance of and adherence to the product prior to any large clinical trials. To us, the concept of a topical microbicide being as convenient and acceptable as a well-tolerated once daily oral tablet, is a high bar to reach.
Two real advances in PrEP pharmacotherapy were the PROUD and Ipergay studies. The PROUD study (Abstract 22LB, webcast) was an open-label RCT of pre-exposure prophylaxis in MSM in England. 545 MSM participants were randomized to receive TDF/FTC either immediately or after a 12-month delay. The deferred PrEP arm was halted prior to study end due to early efficacy results of the immediate treatment arm. A relative risk reduction in HIV acquisition of 86% was seen in the immediate versus delayed treatment arm. There was no evidence of an increase in STI's in the immediately treated population within this study, suggestive of no increase in riskier sexual practices.
The ANRS Ipergay study (Abstract 23LB, webcast) investigated the efficacy of "on-demand" PrEP in high risk MSM. The study enrolled 400 MSM and randomized to either TDF-FTC (300mg/200mg) or placebo. Participants were asked to take two tablets, 2 to 24 hours before each sexual intercourse, then another tablet 24 hours later and a 4th tablet 48 hours after the first drug intake. After a median follow-up of 8.8 months there was an 86% relative risk reduction in the incidence of HIV with the on-demand PrEP (95%CI 39.4-98.5%, P=0.002) in the TDF/FTC arm. One thing to note is that patients used a median of 14 tablets/month, which may bring into question the possibility of residual drug concentrations from previous dosing encounters. Regardless, the strategy was shown to be highly effective in the real world, and we expect additional similar studies to be conducted in the future as PrEP uptake continues to grow.
Jose Castillo-Mancilla presented excellent pharmacology work (Abstract 83, webcast advancing the red blood cell (RBC) matrix measured in a dried blood spot (DBS) as a marker for adherence. We previously knew TFV concentrations in RBC's could be used as a marker of cumulative dosing. Dr. Castillo-Mancilla complimented this knowledge by presenting new data detailing the use of FTC as a marker for recent dosing and adherence in RBC's. The difference in intracellular half-lives of TFV (17 days) and FTC (31 hours) in RBCs is what makes an RBC DBS test attractive for concurrently assessing recent and cumulative dosing in patients receiving TDF/FTC.
Abstract #525 [poster] filled in key pharmacology data with respect to sex dependent differences in drug penetration into various anatomical compartments. The authors found that TFV-DP and FTC-DP display differential penetration at steady state into blood, rectal and genital compartments in HIV positive and negative males and females. Steady state rectal TFV-DP concentrations were ~10 fold higher than concentrations observed in either the genital or blood compartments for either sex. Females appeared to have higher TFV-DP concentrations in the rectal MNC's than males, although the difference wasn't statistically significant. Genital cell concentrations at steady state were ~10 fold higher in females compared with males for both TFV-DP and FTC-DP. Collectively these data describe sex dependent differences in penetration of TFV-DP and FTC-DP into anatomical compartments. These data will help inform the field of dose-response relationships for both prevention and treatment of HIV.
Abstract #966LB [poster] explored the PK-PD relationship of long-acting cabotegravir (CAB LA, GSK1265744) when used as a PrEP agent. The authors assessed plasma PK and longitudinal breakthrough infections in rhesus macaques following multi-exposure intravaginal challenges. Macaques were administered a loading dose (10, 30 or 50mg/kg) of CAB LA on day -7 and a second dose (10, 30 or 50mg/kg) of CAB LA on day 0. These doses were followed by weekly intravaginal SIV challenges. The results showed the 30mg/kg and 50mg/kg dosing strategies observed similar plasma PK, near the predicted plasma PK of the 50mg/kg dose. Significant protection from infection occurred in the macaques dosed at 30mg/kg and 50mg/kg with a median of 9-11 challenges needed to establish an infection. Plasma concentrations of CAB >0.664μg/mL (4x the protein adjusted IC90) yielded 98% protection; 3/140 challenges produced infection at these concentrations. These PK-PD data provide a relationship between plasma drug concentrations and in-vivo efficacy and should inform development of CAB LA as a long-acting PrEP agent.
Finally, a bit of commentary about adherence as it pertains to PrEP. We have now seen poor adherence of study drug drive the failure of multiple PrEP trials. These trials often require vast amounts of resources, as evidenced with the failure of the VOICE study at a price tag of $93 million dollars. Within VOICE patient self reported adherence was measured to be 90% while objective drug concentration analysis depicted adherence levels between 25-30%; a stark difference. Recent efforts have been focused on determining factors for non-adherence. CROI abstract #979 poster] investigated correlates of adherence within the VOICE trial between both the oral and the vaginal PrEP products. The authors found that the correlates differed between the two products. For instance, in the group randomized to the oral product, higher correlates of use were associated with greater amounts of weekly alcohol consumption. Within the group randomized to the gel product the opposite was true, higher rates of drug detection (and thus product use) were seen in women that never used alcohol. This abstract is just one example depicting the need to better understand factors that influence drug adherence. Additionally, there is a need to measure adherence better, in a more frequent and less invasive manner than what is widely used now. These tools will allow for better patient and provider feedback with the goal of avoiding a repetition of the costly trial failures that have occurred in the past. Adherence tools such as the RBC DBS test are becoming very informative and convenient to use. Hopefully we have seen the end of discovering poor adherence in large, expensive clinical trials, after the fact.
III. Drug-Drug Interactions
1. A very clinically significant interaction: efavirenz (EFV) and subdermal levonorgestrel (LNG) contraceptive implant (Abstract 85LB webcast). The LNG implant is a recommended and widely used contraceptive in resource-limited settings. It has an intended duration of use of 4-5 years after insertion, and an expected failure rate of < 1%. LNG is metabolized by CYP3A4, and EFV is a well-known inducer of CYP3A, thus setting up the potential for an interaction that reduces LNG concentrations. This is exactly what Scarsi and her colleagues found. There was approximately a 48% reduction in LNG concentrations in HIV-infected women receiving EFV-containing ART who had the LNG implant inserted (N=20). There were 3 unintended, breakthrough pregnancies in these women, for a contraceptive failure rate of 15% (3/20); there were no breakthrough pregnancies in the control group. There are some very clear implications of this work: women on EFV who have the LNG implant should be counseled regarding the possibility of unintended pregnancies and the addition of other forms of contraception should be considered. Effective contraception for HIV-infected women on EFV-containing ART is a public health priority. More research is urgently needed on ART options that can be used with the LNG implant.
All in all, CROI 2015 was an outstanding meeting, and progress on several fronts was reported. CROI 2016 will be in Boston, MA, February 21-25, 2016 - - and hopefully the Boston winter of 2016 will not be a repeat of the Boston winter of 2015!