Skip directly to content

Policy Statement on Application for Approval of Adefovir for HIV


Policy Statement on NDA 20-993, Gilead Sciences, Inc's Application for Approval of Preveon™ adefovir dipivoxil for Treatment of HIV

1 November 1999

New York, NY


On June 29th, 1999, Gilead Sciences, Inc. submitted new drug application (NDA) 20-993 to the US Food and Drug Administration (FDA) for accelerated approval of Preveon™ brand adefovir dipivoxil (formerly known as bis-pom PMEA). Adefovir is a nucleotide analog that inhibits the reverse transcriptase (RT) enzyme of the human immunodeficiency virus (HIV). Unlike the nucleoside analog reverse transcriptase inhibitors (RTIs), such as AZT, ddI, d4T and 3TC, adefovir, which possesses a phosphate group, requires only two intracellular metabolic steps to be converted into its active form. While intriguing data have been presented regarding the activity of adefovir dipivoxil against HIV that has developed resistance mutations to lamivudine (3TC), the overwhelming preponderance of studies of the drug's activity have been disappointing, and the rate of serious renal toxicity in patients treated with adefovir is high. Consequently, the application for approval of the drug presents regulators and patient advocates with particular challenges. 

TAG's Position

Based on data that have been publicly presented at medical conferences, as well as data provided by the sponsor and on our analysis of the relevant legislative and regulatory standards and guidelines, TAG cannot endorse approval of adefovir at this time, as current data do not indicate that Preveon™, at the dose regimen proposed for approval, is safe or effective according to the accelerated approval regulatory standard.

However, because we recognize that the need for therapies for pre-treated patients is strong, we believe that Gilead should continue to study the drug in this patient population, and should continue to make the drug available to patients lacking other treatment options through expanded access programs.

Moreover, should FDA grant accelerated approval for the current application, recognizing that the need for therapies for pre-treated patients is strong, we believe the approval should be with a very narrow indication focused on heavily pre-treated patients. At the same time, Gilead should continue to study the drug in this patient population, with particular focus on patients who have the lamivudine resistance mutation which preliminary studies have suggested may confer hypersensitivity to adefovir.

TAG would also like to note that, while the therapeutic margin for the use of adefovir against HIV seems to be relatively small, the drug has shown great promise in the treatment of hepatitis B, where the antiviral activity seems to be much greater, allowing for use of lower doses, and potentially reduced incidence of renal toxicity. TAG strongly encourages Gilead Sciences to continue studying the drug as a therapy for hepatitis B.

1. Activity and Efficacy

In vitro, adefovir potently inhibits HIV in a number of cell lines including monocytes and macrophages. Adefovir is synergistic in vitro with d4T, ddC, AZT, nelfinavir, ritonavir, and saquinavir.

In Study GS-408, 442 patients on failing antiretroviral therapy were randomly assigned to add 120 milligrams (mg) of adefovir (ADV) once daily (qd) or placebo to background therapy. Participants had an average baseline CD4 cell count of 355, and an average viral load of 4.4 log10 (henceforth log). After twenty-four weeks of treatment, ADV-treated patients had a significant viral load decline of 0.53 log copies/ml as compared to placebo patients. In addition, a post hoc subset analysis showed that patients with the M184V reverse transcriptase mutation (associated with resistance to lamivudine) achieved an average 0.94 log reduction in plasma HIV RNA levels by week 24 (Jaffe and Miller).

In CPCRA 039, 505 patients were randomly assigned to add 120 mg ADV or placebo to background therapy. At baseline, 85% of patients were taking protease inhibitors, and 26% had plasma viral load counts below 500 copies/ml. The study was intended to be a clinical endpoint study and to enroll many more patients. However, due to the unexpected low rate of clinical events, the study was stopped early, with a median follow-up of eleven months. No significant differences were seen in rates of survival or in viral load reduction. However, due to the premature stopping of this study, a clinical difference was unlikely. The sponsor claims that the expected minimal viral load reduction associated with ADV may have been masked by changes in background therapy (Fisher).

In ACTG 359, a complicated study compared ADV to delavirdine to the combination of the two therapies in a group with indinavir experience and viral load between 2,000-200,000 copies/ml. At the week sixteen primary analysis, delavirdine was superior to adefovir, but the combination of delavirdine and adefovir was no better than adefovir alone. The investigators have hypothesized that the failure of adefovir to add to the virologic improvement of these patients may have been related to a possible pharmacokinetic interaction between adefovir and delavirdine which resulted in lower levels of the NNRTI, though the sponsor has disputed these findings. [By 16 weeks, only 1% of the subjects in the study had developed proximal renal tubular dysfunction (PRTD), though additional cases developed afterwards, as might be expected]. The ACTG 359 protocol team speculated that:

The explanation for the inferior virologic effect in the adefovir arms may be that subjects had extensive nucleoside experience or discontinued lamivudine [3TC] upon study entry... The lack of additivity or synergy in the delavirdine plus adefovir combination arms may well be due to an adverse pharmacokinetic interaction between delavirdine and adefovir first demonstrated in the intensive pharmacokinetic substudy (ACTG 884) of the current study... [which] showed that delavirdine levels were halved when co-administered with adefovir... In addition, saquinavir levels were reduced by about half in the delavirdine plus adefovir combination arms, possibly as a direct result of the decreased delavirdine levels (ACTG 359 team).

In Study GS-420, forty-seven patients were treated with 60 mg ADV or placebo for four weeks to determine the activity of the lower dose of ADV. After four weeks, patients in the ADV-treated group had an average viral load decline of 0.25 log copies/ml, as compared to the placebo-treated group, which had an average 0.08 log increase in plasma viral load (Jaffe).

In study GS-417, 214 RTI experienced, protease naive patients were randomly assigned to receive either 120 mg ADV or 60 mg ADV (each qd) in combination with various nucleoside analogs (ZDV, d4T, 3TC) and protease inhibitors (SQVsgc, NFV). Mean baseline HIV RNA level was 4.6 log copies/ml, and mean CD4 cell count was 362 cells. Patients could have received prior treatment with nucleoside analogs and non-nucleoside analog reverse transcriptase inhibitors; however all patients were naïve to protease inhibitors. At week twenty, there was a mean reduction in HIV RNA levels of 1.2 log for both groups, and a mean increase of about 80 CD4 cells. The proportion of patients with undetectable HIV RNA levels (below 400 copies/ml) was 41% in the 60 mg ADV group, and 31% in the 120 mg ADV group by intent to treat (ITT) analysis; an as treated analysis revealed a smaller difference of 48% BLQ (60 mg) vs. 45% (120 mg) (Gallant 1999). Viral load reductions were comparable in each group; however, the power to detect a difference was small due to the drug's expected small contribution to the overall virologic effect and the study's limited size.

2. Safety

Since a Fanconi-like syndrome known as proximal renal tubular dysfunction (PRTD) emerged as the most serious toxicity of adefovir, safety data on PRTD have been collected from a number of studies. Laboratory signs of PRTD include elevations in serum creatinine levels, reductions in phosphate and sodium bicarbonate levels, proteinurea and glycosurea. Most cases seen were mild to moderate. Onset is slow, normally occurring after about 24 weeks of treatment. Clinically, the illness may result in serious kidney problems and bone density loss. Risk factors for the syndrome include increased age and elevated baseline serum phosphate levels. Non whites appear to be less likely to develop PRTD, according to Gilead's analyses (Jaffe). Typically, the syndrome has been handled clinically by either reducing the dose of adefovir from 120 mg to 60 mg a day, or by discontinuing treatment. Signs and symptoms do seem to resolve following intervention, suggesting that careful monitoring may reduce the danger associated with PRTD.

Adefovir was originally studied primarily at a dose of 120 mg qd. In study GS-408, 1% of patients taking adefovir at this dose had grade three or four PRTD at week 24, however by 48 weeks, approximately 33% of patients were affected with PRTD. 

In CPCRA 039, however, as many as one-fifth of participants had evidence of PRTD during the eleven months of follow-up, almost one-third of patients had some laboratory abnormalities suggesting renal problems, and almost 40% had stopped drug by one year (Gilead, 8 Sept 99 and Fisher 1999). 

In study GS-417, which was designed to assess the relative rates of PRTD for patients treated with 120 vs. 60 mg of adefovir per day, the forty-two week rate of elevated creatinine levels was 50% for the high-dose group, versus 28% for the low-dose group (patients in this study had ADV discontinued when laboratory abnormalities such as elevated creatinine suggested onset of PRTD) (Jaffe). 

In the Gilead Expanded Access program, which had enrolled almost 9,000 patients as of September 1, 1999 -- 5,718 at 120 mg and 3,212 at 60 mg ADV -- the median duration of therapy was six months, so the ultimate incidence of PRTD appeared low. In a snapshot of the expanded access safety database presented at this year's Retrovirus Conference, creatinine elevation and/or proteinuria caused 57/4,519 patients to discontinue study treatment. Gastrointestinal complaints led another 87 patients to withdraw from the program, but other adverse events were relatively rare and non-serious (Nuessle). This low number of renal toxicities, again, may be a function of the limited time on treatment for most of these patients, and may not reflect the overall incidence of Fanconi-like syndrome in patients treated with adefovir for extended periods of time.

At present, the company's expanded access program requires monthly urinalysis and bloodwork to look for symptoms of PRTD. Gilead representatives insist that their educational material will recommend such tests for all patients treated with adefovir if the drug is approved, and has offered to develop a payment mechanism to ensure that indigent patients have access to this vital
monitoring (Jaffe).

3. Resistance

Gilead has conducted a number of in vitro studies to predict what mutations might be associated with decreasing susceptibility to adefovir. In vitro, mutations are seen at codon 70, causing a nine-fold decrease in susceptibility to adefovir, and at position 65, which decreased susceptibility to adefovir by sixteen-fold. This mutation has been observed in 10-15% of patients treated with ddC (Cherrington 1997).

Over long-term follow-up of the ongoing clinical studies, researchers from Gilead have had little success in finding mutations in the reverse transcriptase enzyme in patients treated with adefovir (Miller 1998). In one study, the mutation at position 70 was observed in viral isolates from two of twenty-nine patients on adefovir. Even so, viral load reductions were still seen (Mulato 1998). Virus containing this mutation seems to be less fit than wild type virus, which could explain the continued suppression of viral load (Miller 1998). Only a few other conserved mutations were observed in patients on concurrent therapies, all of whom continued to sustain reductions in viral load. In one study, patients with high-level phenotypic and genotypic resistance to AZT and 3TC, but without the M184V mutation most commonly associated with 3TC resistance, had a significantly reduced response to adefovir therapy (Miller 1999).

4. Drug Interactions

Although few serious interactions were expected between adefovir and other commonly used anti-HIV medications, in ACTG 884 -- a pharmacokinetic substudy of ACTG 359 -- investigators noted a significant, unexpected and deleterious interaction between adefovir and delavirdine. In this study, delavirdine levels decreased significantly in the presence of adefovir. As noted above, these findings are being disputed by the company, and a re-analysis is underway (Acosta).

However, there is a surprising lack of published data regarding adefovir with other commonly used medications. Few interactions are expected, but given the drug's toxicity, and the unexpected findings from ACTG 359, more data would provide more comfort regarding the drugs use in the salvage setting, where polypharmacy is the rule.

5. Pediatrics

Gilead has not filed an application for approval of adefovir in pediatric patients at this time. However, the company is conducting an ongoing study (GS-418) of the drug as part of a HAART regimen in HIV-infected children.

6. Follow-up Studies

According to representatives of Gilead Sciences, the company will continue to study adefovir in GS-415, a therapeutic intensification study, and in GS-458, a study of the drug in patients whose first HAART regimen has failed. In addition, the company has an ongoing pediatric study, and has committed to intensive clinical virology and pharmacokinetic studies with adefovir. Finally, Gilead has committed to follow 2,000 patients for up to five years to assess the long-term impact of adefovir
treatment (Jaffe).

7. Discussion

The adefovir accelerated NDA poses a number of difficult issues. The drug's antiviral activity is far from impressive, especially in pre-treated patients, who are just those for whom the sponsor seeks approval to market the drug. Moreover, the cumulative incidence of renal dysfunction at the more heavily studied 120 mg dose appears to be over one third at one year. The incidence appears to be delayed at the 60 mg dose for which the sponsor seeks approval. However, the database on the 60 mg dose is small and short-term. According to the sponsor, the median time (50% of drug discontinuers) to study drug discontinuation in the randomized study 417 was 26 weeks at the 120 mg dose and 30 weeks at the 60 mg dose. Thus, for those in this study who developed renal dysfunction, the lower dose afforded about four more weeks of antiviral activity before toxicity required drug discontinuation.

In essence, the sponsor is asking the FDA to approve a drug based on antiviral activity demonstrated for the 120 mg qd dose, with safety data based on much shorter follow-up, in smaller numbers of patients, with the 60 mg dose. The activity claims for this dose rest on two-to-four week data from the GS-420 monotherapy study (which showed a <0.4 log reduction at two weeks, with a slight return towards 0.3 log at week 4), and combination efficacy data from the RTI-experienced study GS-417, which demonstrated that 30% of patients in both the 60 and 120 mg arms achieved a viral load below the limit of quantification (400 copies/ml) by week 8, and that this proportion continued until week 20 (rising to 40% in the 60 mg arm). However, the contribution of ADV at either dose in this study is difficult to determine, since all participants also took either SQVsgc/NFV, NFV/1 NRTI or SQVsgc/1 NRTI.

The sponsor disputes negative results produced by independent clinical trials groups, claiming that the CPCRA 039 study (N=503) lacked the power to determine even a viral load difference, let alone a clinical one, and that the pharmacokinetic interaction between DLV and ADV suggested in ACTG 359 -- and shown by PK analysis in substudy 884 -- was spurious. Curiously, then the sponsor proposes a study design similar to CPCRA 039 (HAART intensification with ADV in GS 415) as its confirmatory post-marketing study. Curiously, as well, if -- as the sponsor argues -- there is no pharmacokinetic interaction between DLV and ADV, then lack of additive virologic benefit with DLV+ADV vs. DLV in ACTG 359 suggests that ADV simply lacked efficacy in this pre-treated population.

Basically, then, the sponsor falls back on the argument that 1) there is a demonstrated need for easy to administer (e.g., qd) agents, 2) there is a demonstrated need for new antiretroviral agents in pretreated patients and 3) since the median length of time on therapy in second, third, fourth and fifth line regimens is eight, eight, seven and six months respectively, that the longer-term renal toxicity of ADV is unlikely to compromise clinical efficacy in a pre-treated patient population.

The intense renal monitoring required by ADV reduces its "ease of use." The real-world use of ADV will vary greatly in duration. The longer-term safety and tolerability of ADV at 60 mg remain to be determined. The incidence of PRTD at one year on 60 mg may well be as high as at 24 weeks on 120 mg, if the issue is one of cumulative exposure, rather than peak blood levels. This toxicity lacks clinical symptoms, and thus, missed appointments and missed lab monitoring will result in some people receiving ADV while developing PTRD. We do not know what will happen to these patients when the drug is not stopped in time. The lack of demonstrated efficacy in pre-treated patients fails to support its indication for this population. The effects on 3TC resistant virus in vitro are encouraging, but remain to be validated clinically.

The last time an antiretroviral drug with such equivocal efficacy came to FDA for review was in late 1996, when the agency considered Pharmacia & Upjohn's application for Rescriptor™ brand delavirdine (DLV). TAG, along with other treatment advocacy groups, failed to recommend approval, but cautioned that the lack of sufficient pharmacokinetic interaction data on DLV was likely to complicate its clinical use. In this case, there is a similar lack of PK interaction data B but here, unlike the case with DLV, there is also clear evidence of possibly cumulative, serious to severe, possibly life-threatening toxicity. While this appears to be reversible, in the absence of appropriate close monitoring the use of ADV, even at the 60 mg qd dose, may cause significant harm in the real world setting, especially as its interactions with commonly used HIV and other drugs has yet to be clearly defined.

If it were simply a case of approval now versus stopping development of the drug for good, the case for immediate accelerated approval might be stronger. But there is an alternative B to continue studying the drug while more clearly defining a safe and active dose for HIV infection, as it continues to be studied (at the significantly lower doses of 10 and 30 mg qd) for hepatitis B virus (HBV) infection. In the meantime, the drug should continue to be made accessible to pretreated patients on the expanded access program.

8. Summary

In summary, TAG has concluded that, based on data that have been publicly presented at medical conferences, as well as data provided by the sponsor and on our analysis of the relevant legislative and regulatory standards and guidelines, TAG cannot endorse approval of adefovir at this time, as current data do not indicate that Preveon™, at the dose regimen proposed for approval, is safe or effective according to the accelerated approval regulatory standard.

We note with particular disappointment that the sponsor met in closed session with the FDA Antiviral Drugs Advisory Committee (AVDAC) in July 1998 to determine the process for fast-tracking adefovir for accelerated approval, eligibility for which it received in November 1998. The sponsor also had a pre-NDA meeting with FDA in February 1999. It is clear from our analysis that the sponsor did as little as it could to determine the safety and efficacy of adefovir at the 60 mg daily dose for which it now seeks approval. We believe that the two studies offered here as evidence of "efficacy" -- a four-week monotherapy study (GS 420, N=45) and a small, short-term virologic equivalence study comparing the 60 and 120 mg doses (GS 417, N=214), from which it is impossible to determine the contribution of ADV at either dose B fail to meet even the relatively lenient regulatory standards for accelerated approval of antiretroviral drugs. Were the drug clearly safe for long-term use in a heterogeneous, heavily pre-treated patient population, and were its interactions with other commonly used anti-HIV and other drugs more clearly defined, we might feel differently. However, while the drug's marginal activity is not its fault, the sponsor should bear full responsibility for its anemic development plan. We hope they take a more robust approach to development of tenofovir (PMPA).

Thus, TAG believes that:

  1. FDA should not approve adefovir at this time.
  2. Gilead should continue to study the drug (at the 60 mg and even lower doses) in highly-pre-treated patients. 
  3. Gilead should continue to make the drug available to HIV patients through expanded access.
  4. Should FDA grant accelerated approval to adefovir at this time,
    • It should be indicated only for use in a very narrow group of patients, specifically those with extensive prior anti-HIV therapy and few other treatment options; 
    • FDA should require a physician licensure program, as is done for thalidomide, to ensure that doctors are properly educated about clinical monitoring and management of patients using adefovir, and that regular reporting to FDA of renal function levels should be a mandatory part of its licensure. 
  5. Gilead should continue pediatric studies using adefovir.
  6. Gilead should rapidly conduct interaction studies with most other major anti-HIV drugs, as well as drugs that are frequently used in HIV-infected patients, such as those used to treat and prevent OIs, antidepressants, methadone, anti-seizure medications, antipsychotics, and birth control pills. 

 * * *

Acosta EP et al, Pharmacokinetic (PK) Evaluation of Saquinavir Soft Gel Capsules (SQV)/Ritonavir (RTV) or SQV/Nelfinavir (NFV) in combination with Delavirdine (DLV) and/or Adefovir Dipivoxil (ADV) - - ACTG 359. 6th Conference on Retroviruses and Opportunistic Infections, Chicago, January 30-Feb 2 1999, Abstract 491.

ACTG 359 Team, Activity of the Soft Gelatin Capsule of Saquinavir (SQVsgc) in Combination with Ritonavir or Nelfinavir and Combinations of Delavirdine and/or Adefovir Dipivoxil in HIV-infected Subjects with Prior Indinavir Use and Viral Loads of at Least 2,000 and No More than 200,000 Copies HIV RNA/ml. Executive Summary, 1 June 1999.

Cherrington JM, Mulato AS, Lamy PL et al. Genotypic characterization of HIV-1 variants isolated from AIDS patients treated with adefovir dipivoxil (bis-POM PMEA). [Abstract] 4th Conference on Retroviruses and Opportunistic Infections, Washington, D.C., Jan 22-26, 1997.

Fisher E, Placebo (PLC)-Controlled, Multicenter Trial of Adefovir Dipivoxil (ADV) in Patients (PT) with HIV Disease, 6th Conference on Retroviruses and Opportunistic Infections, Chicago, January 30-Feb 2 1999, Abstract 491.

Gallant J, et al. Randomized, Double-Blind Study of Adefovir Dipivoxil (ADV) at Two Dose Levels in Combination with Nucleoside Reverse Transcriptase Inhibitors (NRTI) and Protease Inhibitors (PI) for NRTI-Experienced, PI-Naïve Patients, 39th ICAAC Program On-Line, accessed 2 Oct 1999.

Jaffe H, Sr. Vice President for Drug Development, Gilead Sciences, Inc., personal communication, 9 Sept 1999.

Miller MD, Antiviral Susceptibilities of HIV-1 RT Recombinant Viruses Derived from AIDS Patients After Prolonged Adefovir Dipivoxil Therapy, 5th Conference on Retroviruses and Opportunistic Infections, January 1998, Abstract 677.

Miller MD et al, Retrospective Analysis of Baseline Susceptability to Adefovir Dipivoxil (ADV) is Predictive of Virologic Response in GS-96-408, 39th ICAAC Program On-Line, accessed 2 Oct 1999.

Miller MD et al, Response to Therapy with Adefovir Dipivoxil is Durable for 48 Weeks and Correlates with Baseline HIV Genotype and In Vitro Susceptability to Adefovir, 6th Conference on Retroviruses and Opportunistic Infections, Chicago, January 30-Feb 2 1999, Abstract 137.

Mulato AS, Genotypic and Phenotypic Characterization of Human Immunodeficiency Virus Type 1 Variants Isolated from AIDS Patients After Prolonged Adefovir Dipivoxil Therapy, Antimicrob Agents Chemother 1998 Jul;42(7):1620-8.

Nuessle SJ et al, The Preveon7 Expanded Access Program: Safety of Adefovir Dipivoxil (ADV) in Antiretroviral Treatment Experienced Patients with Advanced HIV Disease, 6th Conference on Retroviruses and Opportunistic Infections, Chicago, January 30-Feb 2 1999, Abstract 379.

Treatment Action Group, "Treatment Action Group Concerned About Approval of New Anti-HIV Drug: AIDS Activists Say Evidence Unclear on Drug's Safety and Efficacy," 10 March 1997.

* * *