MEMORANDUM

 

DATE:                         April 14, 2003

 

TO:                              Advisory Committee Members and Guests

 

FROM:                        Atazanavir Review Team

 

THROUGH:                 Debra Birnkrant, M.D.

                                    Division Director

                                    Division of Antiviral Drug Products

 

SUBJECT:                   Background Package for NDA 21-567: atazanavir sulfate

 

I.                   Summary of Regulatory Issues and Purpose of Meeting

 

This document provides background information for the May 13, 2003, Antiviral Drugs Advisory committee meeting on atazanavir sulfate. On this day, the committee will be asked to consider efficacy and safety data submitted to support the approval of atazanavir for the treatment of HIV infection.

 

The FDA analyses of the safety and efficacy data submitted in the NDA support the applicant’s findings. Phase 2 and 3 trials submitted in support of this NDA provide evidence that the antiviral activity of atazanavir is similar to nelfinavir or efavirenz in combination with two NRTIs in treatment-naïve patients. In a registrational study of treatment-experienced subjects, atazanavir was inferior to lopinavir/ritonavir both in terms of viral load reduction and percentage of patients with viral load below limits of quantification; however, multiple analyses performed by FDA and the applicant support that atazanavir has antiviral activity in this population.

 

Highly treatment-experienced subjects having failed at least two regimens containing drugs from all three classes were enrolled in study 045. A ritonavir-boosted dose of atazanavir, and atazanavir given in combination with saquinavir were compared to lopinavir/ritonavir, each with tenofovir and an NRTI. Preliminary results support the similarity of the ritonavir-boosted dose of atazanavir to lopinavir/ritonavir, while the ATV/SQV arm appears to be inferior. These data are preliminary and efficacy data from this trial will not be used to make a regulatory decision on this NDA.

 

The Division is convening this meeting to solicit the committee’s comments on the breadth of the proposed treatment indication, and the risk-benefit analysis of the use of atazanavir as it relates to the following safety issues: 1) the incidence and degree of hyperbilirubinemia seen in clinical trials, 2) prolongation of the QT and PR interval, and 3) lipid profiles observed in atazanavir subjects as compared to efavirenz and selected protease inhibitors. Given the diversity of these issues, we have invited several committee guests with expertise in fields relating to these safety issues.

 

The applicant is proposing a broad indication for the treatment of HIV infection based on the results of the two registrational trials. While atazanavir was comparable to currently marketed ARV medications in treatment-naïve studies, it appeared inferior to lopinavir/ritonavir in treatment-experienced patients. We would like to hear comments from the committee regarding the proposed treatment indication for atazanavir.

 

With regard to safety issues, three areas of concern emerged during the atazanavir development program. The first is the frequency of hyperbilirubinemia seen in atazanavir-treated subjects; this adverse event is dose dependent and appears to be due to inhibition of UDP-glucuronosyl transferase, an enzyme responsible for the conjugation of bilirubin. Over three-fourths of all patients experienced an elevation of bilirubin while on treatment, and approximately five percent of patients experienced grade 4 (five x upper limit of normal) increases requiring dose modification per study protocols. Treatment discontinuations for jaundice and/or scleral icterus were uncommon despite a 15% incidence of these events.

 

The hyperbilirubinemia observed in atazanavir-treated subjects was predominantly indirect, regardless of the degree of hyperbilirubinemia observed. Significant elevations of direct bilirubin appeared to occur predominantly in association with other indices of hepatic injury or inflammation. Discontinuations due to abnormal LFTs or hepatotoxicity (lactic acidosis syndrome/symptomatic hyperlactatemia [LAS/SHL] cases were examined separately) appeared to occur with similar frequency between atazanavir and comparator regimens.

 

In two phase 2 studies that compared atazanavir to nelfinavir, each with identical NRTI background therapy, the frequency of all grades of transaminase abnormalities was higher in atazanavir arms. The incidence of grade 3-4 transaminase elevations was higher in atazanavir arms in one of these studies, but lower in atazanavir arms in the second study.

 

In registrational study 034 which compared atazanavir to efavirenz in treatment-naïve patients, the incidence of all grades of transaminase abnormalities was similar between treatment arms. In registrational study 043 of treatment-experienced patients, atazanavir subjects experienced more grade 3-4 LFT abnormalities than lopinavir/ritonavir subjects. Although there was an imbalance in hepatitis B or C co-infection between treatment arms (ATV 20%, LPV/RTV 12%), this did not explain the differences. Slight differences in background NRTI therapy also existed in this study, with use of ddI and d4T being slightly more common in atazanavir subjects.

 

In summary, the hyperbilirubinemia seen during the development program of atazanavir did not appear to result in an increased incidence of hepatotoxicity relative to selected PIs or to efavirenz. During the meeting we will be seeking your assessment of the clinical data with regard to hyperbilirubinemia and the risk of hepatotoxicity associated with atazanavir use. We would like your general impression of the clinical implications of these data and your recommendations for additional preclinical or clinical studies to address the potential for hepatotoxicity.

 

The second safety issue relates to effects of atazanavir on the QT and PR interval. Effects of drugs on the QT interval have become an increasing focus of the FDA; QT prolongation and the subsequent development of Torsades de Pointes (TdP) have been one of the most common reasons for drug withdrawal in recent years. While the risk-benefit analysis of taking an antiretroviral medication versus the possibility of developing an extremely rare but potentially life-threatening arrhythmia may appear to be clear-cut, the Division believes that we have moved into management of HIV infection as a chronic disease. As such, all risks associated with medication use should be well delineated.

 

Evaluation of the QT interval includes “correction” of the QT interval for heart rate, as the QT interval decreases with increasing heart rate. In this document, corrected QT intervals (QTc) were derived using a correction formula known as Bazett’s; this has been the correction method historically used by the FDA and the one on which criteria for evaluation of the QT interval have been based. Evaluation of the QT interval is a specialized and evolving field and will not be discussed at length in this document; further information will be provided at the Advisory Committee meeting with the goal of allowing attendees to participate in a discussion of the QT effects of atazanavir.

 

In brief, a placebo-controlled pharmacokinetic study designed to evaluate effects of atazanavir on ECG parameters revealed a dose-dependent prolongation of the QT interval. Prolongation that may be considered a signal for increased risk for development of TdP was seen at a dose of 800 mg given once daily. This dose produced an exposure that is three-fold greater than that seen with the proposed dose of 400 mg. The ritonavir boosted dose of atazanavir 300 mg that is being investigated for use in treatment-experienced patients has not been fully evaluated in a placebo-controlled pharmacokinetic study, but data suggest that this dose may also be associated with prolongation of the QT interval.

 

In order to further evaluate cardiac risks, ECGs were collected from five clinical trials. Use of atazanavir did not appear to result in an increased incidence of QTc interval prolongation relative to comparators. There were no clinical events of sudden death, or report of arrhythmias that appeared to be related to prolongation of the QT interval; however, these types of clinical events are rare, and likely would not be seen in clinical trials of the size seen in this application.

 

During evaluation of the effects of atazanavir on the QT interval it was also found that atazanavir produced dose-dependent prolongation of the PR interval. The incidence of first degree AV block was common and occurred in over 50% of subjects receiving 800 mg of atazanavir.

 

In clinical trials of atazanavir first degree AV block was observed with similar frequency in atazanavir subjects versus PI comparators. First degree AV block appeared to be less common in subjects receiving efavirenz. In study 034 bundle branch block was reported in one ATV subject and one EFV subject. In the expanded access protocol a patient taking atazanavir concomitantly with verapamil, delavirdine, and other medications, was hospitalized with angina and a junctional rhythm.

 

In summary, while pharmacokinetic studies revealed moderate effects of atazanavir on the PR interval, clinical events related to prolongation of the PR interval were rare. First degree AV block was the most common abnormality observed. Effects on the QT interval at the proposed dose appeared to be minimal. We will be seeking comments from the Advisory Committee on the risk-benefit analysis of the use of atazanavir with regard to these issues.

 

The final safety issue relates to lipid metabolism. It was noted during phase 2 studies of treatment-naïve subjects that treatment with nelfinavir resulted in greater increases in lipid parameters relative to atazanavir. These studies were not designed specifically to measure these changes; however, this finding was confirmed in phase 3 studies of both treatment-naïve and treatment-experienced patients.

 

The applicant analyzed lipid data from all studies in multiple ways. Mean changes from baseline were calculated and categorical analyses were performed using NCEP guidelines to define categories of lipid elevation. Data regarding initiation of lipid-lowering agents during studies were recorded. Calculations were made using Last Observation Carried Forward (LOCF) for subjects initiating lipid lowering therapy during the trial and sensitivity analyses were performed without using LOCF.

 

In general, atazanavir produced less change in total cholesterol, fasting LDL, and triglycerides than all comparators; these differences were found to be statistically significant. Atazanavir subjects initiated lipid lowering therapy less frequently than patients on comparator regimens. After 72 weeks of nelfinavir therapy, lipid levels of subjects who switched from nelfinavir to atazanavir returned to pretreatment levels.

 

Two concerns have emerged with regard to lipid parameters. The first is whether this finding will be maintained over longer durations of therapy and across multiple treatment regimens. In study 043 fasting triglycerides did not appear to decrease significantly in the atazanavir treatment arm to what may be considered pre-treatment (or treatment naive) levels. This may suggest that other factors in addition to protease inhibitor use may contribute to hypertriglyceridemia. The other concern is whether this apparent lack of effect on lipid parameters will translate into health benefits for patients in terms of a lower incidence of lipodystrophy and cardiovascular disease. Spontaneous reporting of lipodystrophy events in these clinical trials does not suggest a reduction of these events in subjects taking atazanavir.

 

We will be seeking the committee’s comments on this potential treatment advantage of atazanavir.

 

 

 

 

 

II.        Clinical Development Summary

 

This NDA contains clinical data collected primarily from nine clinical studies, including the two registrational studies, AI424034 (034) and AI424043 (043). Study 034 was an international, multi-center, double-blind, randomized, placebo-controlled trial comparing atazanavir to efavirenz, each given with AZT/3TC, in treatment-naïve HIV-infected subjects. Study 043 was an international, multi-center, randomized, open-label trial comparing atazanavir to lopinavir/ritonavir, each with an optimized NRTI background, in HIV-infected subjects who had failed a PI-containing regimen.

 

Several supportive studies were also submitted, including studies AI424007 (007) and AI424008 (008), two dose-finding studies comparing atazanavir to nelfinavir. Studies AI434041 (041) and AI424044 (044) were rollover studies for the dose-finding studies and were designed to collect long-term safety data. Also notable is study AI424045 (045), a multi-center, randomized open-label trial comparing a ritonavir-boosted dose of atazanavir, and atazanavir given in combination with saquinavir, to lopinavir/ritonavir, each with tenofovir and an NRTI, in highly treatment-experienced HIV-infected subjects who had failed at least two regimens containing ARV medications from all three classes.

 

Other trials include a PACTG pediatric protocol (020), an expanded access protocol (900), and a small phase 2 trial of treatment-experienced patients (009).

 

Summaries of these trials are provided in the table presented on the following page:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Summary of Clinical Trials

Study

Design

Regimens

     (mg)

Comparator

      (mg)

Background

#

Enrolled

Pt Population

Endpoint

007

Randomized Blinded to ATV dose

ATV 200

         400

         500

Nelfinavir

750 tid

ddI/d4T

420

Treatment naive

TAD*  in log10 HIV RNA D from B/L

008

Randomized

Blinded to ATV dose

ATV 400

         600

Nelfinavir

1250 bid

d4T/3TC

467

Treatment naive

TAD

009

Randomized

ATV 400

SQV 1200

ATV 600

SQV 1200

RTV 400

SQV 400

Optimized background

85

Treatment

experienced

TAD

041

Rollover study for 007 and 009 to collect long-term safety data

ATV 400

NFV 750 tid

Background therapy received in previous trial

222

Subjects completing 007 and 009

Collection of long-term safety data

044

Rollover study for 008 to collect additional safety data

ATV 400

Patients receiving NFV in 008 switched to ATV to assess lipids

Background therapy received in previous study

346

Subjects completing study 008

Collection of long-term safety data

034

Randomized

Double-blind

Placebo controlled

ATV 400

EFV 600 mg

AZT/3TC

810

Treatment

naïve

Percent BLQ

043

Randomized

Open-label

ATV 400

LPV/RTV

Optimized background of 2 NRTIs

300

Patients who failed a PI regimen

TAD 

045

Randomized

Open-label

ATV 300 RTV 100

 

ATV 400

SQV 1200

LPV/RTV

Tenofovir and 1 NRTI based on results of phenotypic testing

358

Highly treatment experienced patients having failed drugs in all three classes

TAD 

900

Expanded Access Protocol

ATV 400

 

ATV 300

RTV 100

None

Based on physician choice

 

Open enrollment

None

020

Pediatric

ATV dose ranging

None

Based on MD choice

43

Age 3 mo to 21 years

PK/PD and safety

*TAD – Time-averaged difference from baseline

 

III.               Summary of Efficacy

 

A.     Dose Selection

 

A dose of 400 mg was chosen based on results from phase 2 dose-ranging studies 007 and 008. No significant differences in efficacy were seen after 48 weeks of treatment with 200 mg, 400 mg, 500 mg, and 600 mg doses of atazanavir; however, an initial two-week monotherapy treatment phase with atazanavir showed that doses of 400 mg or greater had higher probabilities of producing a 1.5 log10 reduction from baseline. The choice of 400 mg provided a balance between efficacy and the incidence of hyperbilirubinemia.

 

B.     Study Design and Baseline Demographics for Registrational Trials

 

As mentioned previously, study 034 was an international, multi-center, double-blind, randomized, placebo-controlled trial comparing atazanavir to efavirenz, each given with AZT/3TC, in treatment-naïve HIV-infected subjects. Study 043 was an international, multi-center, randomized, open-label trial comparing atazanavir to lopinavir/ritonavir, each with an optimized NRTI background, in HIV-infected subjects who had failed a PI containing regimen.

 

Baseline characteristics of subjects enrolled in these studies are summarized on the following page.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Baseline Characteristics: Studies 034 and 043

 

 

Study 034

Study 043

# of Subjects Randomized

810

300

# of Subjects Treated

805

290

Age (Years)

   Mean

   Median

   Range

 

34

33

18, 73

 

38

37

20, 65

Sex (%)

   Male

   Female

 

65

35

 

79

21

Race (%)

   Caucasian

   Hispanic

   Black

   Asian/Other

 

33

37

13

17

 

41

52

7

<1

CD4 Cell Count (cells/mm3)

   Mean

   Median

 

 

322

282

 

 

320

268

HIV RNA

(log10 copies/mL)

   Mean

   Median

 

N < 100,000 (%)

N ³ 100,000 (%)

 

 

4.84

4.88

 

58

42

 

 

4.14

4.19

 

83

17

Mean Time on Prior Antiretroviral Therapy

(weeks)

 

PIs

NRTIs

NNRTIs

 

 

 

 

N/A

 

 

 

 

144 (100% of subjects)

184 (100% of subjects)

94 (14% of subjects)

 

 

C.     Primary Efficacy Endpoints

 

The primary efficacy endpoint in study 034 was percentage of patients with HIV RNA levels below the limit of quantification of 400 copies/mL at 48 weeks. The primary efficacy endpoint for study 043 was the magnitude of viral suppression as assessed by the change from baseline in plasma HIV RNA levels (expressed in log10) through 24 weeks. Multiple secondary analyses were performed for each study.

 

 

D.    HIV RNA Results

 

The following two tables summarize efficacy results for selected trials. The first table provides efficacy results for atazanavir 400 mg in studies 007, 008, and 034. In these studies, atazanavir was similar to efavirenz and nelfinavir in a Time to Loss of Virologic Response (TLOVR) analysis using both 400 copies/mL and 50 copies/mL as limits of detection.

 

At 24 weeks in study 043, subjects receiving atazanavir had a mean decrease of 1.73 log10 c/mL as compared to a mean decrease of 2.16 log10 copies/mL for lopinavir/ritonavir patients. The time-averaged difference (TAD) estimate (ATV - LPV/RTV) for the change from baseline in HIV RNA level through 24 weeks was 0.31 log10 c/mL (97.5% CI: 0.06, 0.55), favoring lopinavir/ritonavir.

 

Preliminary efficacy results at 16 weeks of a limited number of enrolled subjects in study 045 were provided in this NDA. A ritonavir-boosted dose of atazanavir 300 mg appeared to be similar to LPV/RTV, each given with tenofovir and an optimized NRTI. Atazanavir given in combination with saquinavir appeared to be inferior.

 

Time to Loss of Virologic Response (TLOVR)

The TLOVR analysis is an intent-to-treat analysis that examines endpoints using the following definitions of treatment failure for patients who have achieved HIV RNA levels below the limit of quantification:

 

For all subjects with confirmed HIV RNA levels below an assay limit, the time to failure is the earliest time when a specific event had occurred. These events are

·        Death

·        Permanent discontinuation of the study drug or loss to follow-up

·        Introduction of a new ARV drug (unless a background drug is changed for reasons of toxicity or intolerance that are clearly attributable to that drug)

·        Confirmed HIV RNA levels above or equal to an assay

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Summary of Efficacy Treatment-Naïve Studies

Time to Loss of Virologic Response (TLOVR)

 

Study 034

Study 007

Study 008

HIV RNA

ATV

AZT/3TC

 

EFV

AZT/3TC

 

ATV

ddI/d4T

 

NLF

ddI/d4T

ATV

d4T/3TC

NLF

d4T/3TC

 

Number of Subjects/Total (%)

 

 

 

 

 

 

 

< 400 copies/mL

281/404 (70)

258/401 (64)

48/78

(62)

50/82

(61)

123/181

(68)

54/91

(59)

< 50 copies/mL

131/404 (32)

150/401 (37)

26/78

(33)

23/82

(28)

60/181

(33)

35/91

(38)

TAD48

-2.67

-2.74

-2.42

-2.33

-2.51