Share this article on:

The NEAT Study: A 48-Week Open-Label Study to Compare the Antiviral Efficacy and Safety of GW433908 Versus Nelfinavir in Antiretroviral Therapy–Naive HIV-1-Infected Patients

Rodriguez-French, Amalia MD*; Boghossian, Jack MD; Gray, Glenda E. MBBCH, FCP(SA); Nadler, Jeffrey P. MD§; Quinones, Arnaldo R. MD; Sepulveda, Gladys E. MD**; Millard, Judith M. PhD††; Wannamaker, Paul G. BA††

JAIDS Journal of Acquired Immune Deficiency Syndromes: January 1st, 2004 - Volume 35 - Issue 1 - p 22-32
Clinical Science

Objective To compare the efficacy, durability, and tolerability of GW433908 (908), 1400 mg twice-daily (BID), with nelfinavir (NFV), 1250 mg BID.

Methods This was an international, multicenter, randomized, open-label study (NEAT) in antiretroviral therapy (ART)-naive HIV-infected adults with plasma HIV-1 RNA (vRNA) at screening ≥5000 copies/mL (c/mL). Patients were randomly assigned to 908 or NFV (2:1) for a minimum of 48 weeks, with a background of abacavir (ABC) and lamivudine (3TC).

Results A total of 166 patients received randomized treatment with 908 BID and 83 received NFV BID. The population was diverse with regard to race and gender (76% Hispanics and blacks, 31% female) and had advanced HIV disease at screening (45% had vRNA >100,000 c/mL, 48% had CD4 cell counts <200 cells/mm3, 20% had a history of Centers for Disease Control class C events). After 48 weeks of study by an intention-to-treat rebound or discontinuation = failure analysis, a greater proportion of patients in the 908 BID group (66%) than the NFV BID group (51%) achieved vRNA <400 c/mL. Furthermore, more patients with screening vRNA >100,000 c/mL (67 vs. 35%) or CD4 <50 cells/mm3 (48 vs. 24%) achieved undetectable viral loads taking 908 BID compared with NFV BID, respectively. Favorable immunologic responses were observed for both groups. Diarrhea, which was more common in the NFV BID group (18 vs. 5%), was the only drug-related grade 2–4 adverse event with a significant difference (P = 0.002) in incidence between groups.

Conclusion Administration of 908 BID resulted in a potent and sustained antiretroviral response, notably in ART-naive patients with advanced HIV disease. GW433908 was generally well tolerated and provides a convenient dosing option without food or fluid restrictions.

From *Centro de Vacunas y Tratamiento de Enfermedades Febriles, Centro Especializado y Hospital Pediátrico San Fernando, La Sabana, Panama; †Peter Ho Clinic, Saint Michael's Medical Center, Newark, NJ; ‡Perinatal HIV Research Unit, University of the Witwatersrand, Chris Hani Baragwanath Hospital, Soweto, Gauteng, South Africa; §University of South Florida, Hillsborough County Health Department, Tampa, FL; ¶Community Health Care/Center One, Inc., Fort Lauderdale, FL; **Hospital San Lucas, Clinica de Immunologia, Ponce, Puerto Rico; and ††GlaxoSmithKline, Research Triangle Park, NC.

Supported by GlaxoSmithKline Research and Development.

The results of this study were previously presented in part at the 43rd Interscience Conference on Antimicrobial Agents and Chemotherapy (ICAAC), September 27–30, 2002, San Diego, CA, poster H-166, and the 10th Conference on Retroviruses and Opportunistic Infections, February 10–14, 2003, Boston, MA, oral abstract G1-29.

Received for publication June 6, 2003; accepted October 7, 2003.

Reprints: Jeffrey P. Nadler, University of South Florida, Hillsborough County Health Department, 1105 E. Kennedy Boulevard, Tampa, FL 33602-3511 (e-mail:

GW433908 (908, fosamprenavir) is an HIV protease inhibitor (PI) currently in late-stage clinical development for the treatment of HIV infection in adults and children. 1,2 GW433908, formulated as 700-mg film-coated tablets, can be taken without food or water restrictions, may be administered with ritonavir in either a twice-daily (BID) or once-daily (QD) dosing regimen, or alone as part of a BID regimen. 3,4 Upon absorption, 908 is rapidly hydrolyzed to amprenavir plus inorganic phosphate, thus retaining the distinct resistance profile of the active moiety, amprenavir. 5,6

GW433908 has been studied in a phase 2 development program including a dose-ranging study in which 1395 mg (3 × 465-mg tablets) and 1860 mg (4 × 465-mg tablets) BID doses of 908 were shown to be comparable to the standard dose of amprenavir (when administered as Agenerase, 1200 mg BID; GlaxoSmithKline, Ware, UK). Agenerase is formulated as soft-gel capsules containing 150 mg of amprenavir requiring 8 capsules BID. GW433908 achieved a reduction in plasma HIV-1 RNA (vRNA) of approximately 2 log10 copies/mL (c/mL) over 4 weeks while maintaining a favorable tolerability profile. 2 Based on the similarity in plasma amprenavir pharmacokinetics, short-term virologic and immunologic efficacy, and safety profile between the 2 dosage regimens, the 1395-mg BID regimen was selected for further study. Study APV10006 demonstrated bioequivalence between the 465-mg and 700-mg tablet presentations. 7 Newly formulated 700-mg film-coated tablets were used without ritonavir in the NEAT study to achieve a dosing regimen requiring just 2 tablets BID.

In the NEAT study, we compared the efficacy, safety, and tolerability of 908 1400 mg BID, with that of nelfinavir (NFV) 1250 mg BID, over 48 weeks of dosing. Both PIs were given in combination with an abacavir (ABC)/lamivudine (3TC) background. The primary objective was to compare the efficacy and durability of the antiviral response to 908 and NFV when each drug was administered to antiretroviral therapy (ART)–naive subjects in combination with ABC and 3TC. Secondary objectives were to compare the safety and tolerability of 908 and NFV, to compare virologic and immunologic responses, to compare the occurrence of events related to metabolic abnormalities, and to assess the development of viral resistance to the two treatment regimens.

Back to Top | Article Outline



ARV-naive HIV-infected males and females who were at least 13 years of age (or 18 years of age according to local requirements) with plasma HIV-1 RNA (vRNA) ≥5000 c/mL were recruited for this study. “ART-naïve” was defined as having had <4 weeks of therapy with any nucleoside reverse transcriptase inhibitor (NRTI) and never having received a nonnucleoside reverse transcriptase inhibitor (NNRTI) or PI. Patients provided written informed consent to participate in the study. All females were of nonchildbearing potential or tested negative for pregnancy and agreed to use a proven barrier method of contraception.

Patients were excluded if they had medical conditions that could have compromised their safety or interfered with drug absorption; had a history of clinically relevant pancreatitis or hepatitis within the last 6 months; were pregnant or lactating women; were using excluded medications; had treatment with radiation therapy or cytotoxic chemotherapeutic agents within 28 days of study drug administration; or had protocol-listed abnormalities in screening laboratory values, including grade 3 or 4 alanine or aspartate aminotransferase (ALT or AST) levels within 28 days prior to study entry. Patients with an active Centers for Disease Control and Prevention (CDC; 1993 classification) 8 category C event were also excluded, although patients with a history of a category C event were eligible for enrollment. There were no exclusion criteria based on CD4 cell counts.

Back to Top | Article Outline

Study Design

This was an international, randomized, open-label, 2-arm study in ART-naive patients (protocol number: APV30001; NEAT). The study was conducted in the United States, Puerto Rico, Panama, and South Africa. A national, regional, or investigational center ethics committee or institutional review board approved the study protocol at each site. The study was conducted in accordance with Good Clinical Practice and the Declaration of Helsinki (October 1996). The primary endpoint was the proportion of patients with vRNA levels <400 c/mL at 48 weeks. Secondary endpoints included the proportion of patients with vRNA <50 c/mL, absolute and time-averaged (average area under the curve minus baseline, AAUCMB) changes from baseline in vRNA and CD4 cell counts, and differences between treatment groups with respect to the incidence of adverse events, metabolic, and laboratory abnormalities. The emergence of viral resistance was also evaluated.

Patients who were successfully screened for eligibility underwent baseline assessments on day 1 and attended study visits at weeks 1, 2, 4, 8, 12, 16, 20, and 24 and every 8 weeks thereafter until the last randomized patient completed their week 48 study visit. Patients were randomly assigned to receive one of 2 treatments: either 908 1400 mg BID (GlaxoSmithKline, Ware, UK) or NFV 1250 mg BID (Pfizer Pharmaceuticals, New York, NY). Both treatments were administered in combination with ABC (300 mg BID) and 3TC (150 mg BID). A 2:1 randomization of 908 to NFV was used, providing for a larger sample size on the investigational drug to generate a robust safety database for unboosted 908. Nelfinavir was chosen as the comparator in this study as it was the most widely used PI in ART-naive patients at the time the study was initiated. The randomization code was stratified according to screening vRNA concentrations (5000–10,000 c/mL; >10,000–100,000 c/mL; or >100,000 c/mL) and allocated in blocks by investigational site. Central randomization procedures were adopted to allocate treatment of each subject using an interactive voice response system. GW433908 was administered as two 700-mg tablets BID, and NFV was administered as five 250-mg tablets BID. Patients were advised of the NFV dosing instructions that stated that NFV should be administered with a light meal or snack; there were no food or fluid restrictions for 908, ABC, or 3TC. Patients who needed to alter their treatment regimen by changing their PI or adding an NNRTI were required to discontinue from the study. However, patients who were intolerant of either of their background NRTIs could change to another approved NRTI and continue in the study.

Back to Top | Article Outline

Efficacy Assessments

Patients provided blood samples for analysis of vRNA and lymphocyte cell counts at the screening visit (within 28 days of study entry), day 1 (predose), and at weeks 1, 2, 4, 8, 12, 16, 20, and 24 and every 8 weeks thereafter. Blood samples were collected, processed, stored, and shipped according to instructions provided in a central laboratory manual used by all study centers. Plasma HIV-1 RNA was measured using the Roche Amplicor HIV-1 Ultrasensitive Monitor test (version 1.5, ultrasensitive limit of quantification = 50 c/mL). Samples with vRNA >75,000 c/mL were retested using the Roche Amplicor HIV-1 Monitor test (version 1.5, standard assay, limit of quantification = 400 copies/mL). CD4 cell counts were conducted using flow cytometry. Patients were assessed for any new or recurring HIV-associated conditions at each visit.

Back to Top | Article Outline

Safety Assessments

Adverse events were recorded throughout the study, although any conditions clearly related to HIV disease (except lymphoma and invasive cervical carcinoma) were excluded from the definition of an adverse event. Rather, these were recorded specifically as HIV-associated conditions. Abnormal laboratory findings (including hematology, clinical chemistry, and urinalysis) or other abnormal assessments that the investigator judged to be clinically significant were recorded as adverse events. All adverse events were assessed for seriousness, intensity, and relationship to study drug. Drug-related adverse events were reported as related to the entire study drug regimen: either 908 plus ABC/3TC, or NFV plus ABC/3TC. In addition to standard clinical chemistry assessments throughout the study, fasting triglycerides, glucose, total cholesterol, and high-density lipoprotein (HDL)/low-density lipoprotein (LDL) cholesterol were assessed at day 1 and weeks 8, 16, 24, and 48.

Back to Top | Article Outline

Statistical Methods

The planned sample size allowed for 210 patients to be randomly allocated in a 2:1 ratio of 908 BID-NFV BID. The primary purpose of the study was to obtain a robust safety database on the use of unboosted 908, hence the inclusion of twice as many patients enrolled on 908 compared with NFV. Although the primary endpoint was the proportion of patients achieving vRNA <400 c/mL, the study was powered on a secondary endpoint, AAUCMB in vRNA, to test the noninferiority of 908 BID to NFV BID. Using a noninferiority margin of 0.5 log10 copies/mL, and assuming an SD in vRNA of 0.7 log10 copies/mL, the planned number of patients provided approximately 99% power to test the noninferiority of 908 BID to NFV BID at the 2.5% significance level. The sample size required to meet this secondary endpoint allowed for the examination of the stated clinical objective of exploring the safety profile of unboosted 908.

The primary population for the efficacy analyses, the intent-to-treat (ITT) population, included all patients randomized to the study who received at least 1 dose of randomized treatment. The primary strategy for ITT efficacy analyses of proportion endpoints was the rebound/discontinuation = failure (RD = F) analysis where positive responders or “successes” were patients with confirmed vRNA levels <400 c/mL. Treatment failures were defined as those subjects whose vRNA never went <400 c/mL, those subjects who had a confirmed vRNA rebound to >400 c/mL, or those who discontinued randomized PI for any reason. In addition, the primary efficacy endpoint was also analyzed using a missing/discontinuation = failure (M = F) analysis. In this analysis, treatment failures included patients who had vRNA >400 c/mL, or who prematurely discontinued randomized PI, or who had missing data. An ITT observed strategy was employed for analysis of AAUCMB and CD4 cell count. Only subjects with data at a visit were included in the observed analysis, and no missing values were imputed. An additional analysis method, the “per-protocol” analysis, included only patient data collected during treatment with randomized therapy and excluded data following major protocol deviations (e.g., nonadherence, absence of postbaseline data, prohibited concomitant medications, and prior ART history). For all analyses, subjects who switched background NRTIs for reasons of intolerance rather than efficacy were not considered as failures. Changes from baseline in vRNA and CD4 cell counts were reported using observed analyses, in which only subjects with data at a visit were included, and no missing values were imputed.

Treatment differences with respect to the primary endpoint were evaluated statistically with confidence intervals (CIs). In the primary efficacy analysis, the proportion of patients with vRNA <400 c/mL was summarized at each visit and the 2-sided 95% CI around the difference in proportions between the 908 BID group and the NFV BID group at 48 weeks was calculated. The CI was stratified to reflect the randomization strata (screening vRNA ≥5000–10,000 c/mL; >10,000–100,000 c/mL; or >100,000 c/mL) using Mantel-Haenszel weights. For the secondary endpoint (AAUCMB), which was used to test for noninferiority, the two-sided 95% CI around the difference in mean AAUCMB between the 908 BID group and the NFV BID group was calculated. The upper limit was then compared with 0.5 log10 c/mL vRNA to assess noninferiority. The CI was calculated using analysis of variance adjusting for randomization strata.

For the safety analyses, the safety population consisted of all subjects who were randomized into the study with documented evidence of having received at least 1 dose of randomized treatment. All analyses were performed according to the treatment actually received regardless of treatment randomization. Fisher exact tests were performed to compare adverse event incidence rates between treatment groups, where comparisons with P values of <0.01 were considered significant. For selected laboratory parameters (triglycerides, HDL, LDL, and total cholesterol, fasting glucose, AST, ALT, and lipase), a Wilcoxon rank sum test was performed to compare the median change from baseline between the treatment groups at week 48, and P values of <0.01 for these comparisons were also considered significant. In addition to analysis of lipid data by toxicity grade, mean values, and change from baseline, lipid data were also analyzed according to National Cholesterol Education Program (NCEP) Lipid Categories. 9

Back to Top | Article Outline


Patient Disposition and Baseline Characteristics

Of 341 patients screened, 251 patients met eligibility criteria and were randomly allocated between November 22, 2000 and July 20, 2001. Of those, 249 received treatment: 166 patients in the 908 BID group and 83 patients in the NFV BID group. These 249 patients comprised the ITT and safety populations. The per-protocol population included 223 patients: 150 patients in the 908 BID group and 73 patients in the NFV BID group. Reasons for exclusion from the per-protocol population for the 908 BID group and the NFV BID group, respectively, included evidence of nonadherence (12 patients, 4 patients), no postbaseline on-treatment data (3 patients, 4 patients), prohibited concomitant medication (1 patient, 2 patients), and not being ART naive at entry (1 patient in NFV BID group). One patient in the NFV BID group had 2 deviations as reasons for exclusion. All patients who were excluded from the per-protocol population were selected in a blinded fashion.

The study population generally had advanced HIV disease at baseline. The median baseline CD4 cell count was 212 cells/mm3. Furthermore, 48% of patients had CD4 cell counts <200 cells/mm3 and 18% had CD4 cell counts <50 cells/mm3 at baseline. The median baseline vRNA was 4.82 log10 c/mL in the 908 BID group and 4.85 log10 copies/mL in the NFV BID group. Forty-five percent of patients had baseline vRNA >100,000 c/mL and 20% of patients had a CDC class C (AIDS) classification at study entry (Table 1). The study population was also ethnically diverse and included a relatively high proportion of females (31%). Baseline characteristics, including incidence of coinfection with hepatitis, were generally similar between the 2 treatment groups (Table 1).



At the end of the study, a greater proportion of subjects had prematurely discontinued their randomized PI in the NFV BID group than in the 908 BID group (46 vs. 30%, respectively). The reasons for premature discontinuation of the randomized PI were similar between treatment groups, with the exception that more than twice the proportion of subjects who received NFV BID were withdrawn from the randomized PI at the investigators' discretion due to insufficient viral load response (NFV BID: 19%; 908 BID: 7%). Other reasons for premature discontinuation of randomized PI in the 908 BID and NFV BID groups, respectively, included adverse event (5%, 7%), consent withdrawn (4%, 4%), lost to follow-up (10%, 8%), clinical progressions (<1%, 1%), and other (2%, 6%). A total of 17/166 patients (10%) in the 908 BID group and 6/83 (7%) in the NFV BID group changed their background NRTI for intolerance reasons and continued in the study.

Back to Top | Article Outline

Efficacy Analyses

Plasma HIV-1 RNA Concentrations

At week 48, the proportion of patients achieving vRNA <400 c/mL was greater in the 908 BID group than the NFV BID group: 66% (109/166) of patients in the 908 BID group and 51% (42/83) of patients in the NFV BID group (ITT RD = F) achieved this endpoint (Fig. 1). Adjusting for randomization strata (screening HIV-1 RNA ≥5000–10,000 c/mL, >10,000–100,000 c/mL, or >100,000 c/mL) produced a stratified difference of 15% (95% CI: 2–28%). Additionally, more patients experienced virologic failure in the NFV BID group than the 908 BID group: 28% (23/83) vs. 14% (24/166) of patients, respectively. Table 2 defines patient outcomes at the end of the 48-week treatment period.





For the ITT M = F analysis, the proportion of patients achieving vRNA <400 c/mL was also greater in the 908 BID group than the NFV BID group at week 48: 66% (109/166) of patients in the 908 BID group and 48% (40/83) of patients in the NFV BID group. The stratified treatment difference was 17% (95% CI: 5–30%) after adjusting for randomization strata. The per-protocol analysis showed that similar proportions of patients achieved vRNA <400 c/mL at week 48: 94% (104/111) of patients in the 908 BID group and 95% (40/42) of patients in the NFV BID group (stratified treatment difference: −1%; 95% CI: −10 to 7%).

Using the ITT RD = F analysis for proportion of patients achieving vRNA <50 c/mL, 55% (92/166) of patients in the 908 BID group vs. 41% (34/83) of patients in the NFV BID group achieved this endpoint at week 48 (Fig. 1). The stratified difference in proportions was 14% (95% CI: 2–27%) after adjusting for baseline strata. Similarly, the proportion of patients achieving vRNA <50 c/mL with the ITT M = F analysis was 58% in the 908 BID group and 42% in the NFV BID group. The stratified difference in proportions was 16% (95% CI: 3–28%) after adjusting for baseline strata. In the per-protocol analysis, 84% (93/111) of patients in the 908 BID group and 83% (35/42) of patients in the NFV BID group achieved vRNA <50 c/mL at week 48. In this case, the stratified difference between groups was 1% (95% CI: −12 to 14%) after adjusting for baseline strata.

In patients who entered the study with high vRNA (>100,000 c/mL), more patients in the 908 BID group than the NFV BID group achieved vRNA below the limits of detection using both the <400 and <50 c/mL endpoints. At week 48, 67% (49/73) of 908 BID patients entering the study with high viral load compared with 35% (13/37) of NFV BID patients achieved vRNA <400 c/mL and 55% (40/73) of 908 BID patients compared with 24% (9/37) of NFV BID patients achieved vRNA <50 c/mL (ITT RD = F, Fig. 2A and B). As noted above, the patient population enrolled into the NEAT study had relatively advanced HIV disease, including 16% of patients who had screening vRNA >500,000 c/mL. Similar to the analysis in subjects with entry vRNA >100,000 c/mL, an analysis at week 48 of those subjects with vRNA >500,000 c/mL at screening indicated that 65% (15/23) of 908 BID patients and 31% (5/16) of NFV BID patients achieved <400 c/mL at week 48. Patients who entered the study with CD4 cell counts <200 cells/mm3 achieved vRNA <400 c/mL at 48 weeks in greater proportions on 908 BID than on NFV BID (68 vs. 41%, ITT RD = F). A similar trend was seen in patients who entered the study with CD4 cell counts <50 cells/mm3 (48 vs. 24% for 908 BID and NFV BID, respectively, ITT RD = F) (Fig. 2C).



A 2 log10 c/mL decrease in vRNA was observed by week 4, with continued decreases to a 3 log10 c/mL reduction that was then sustained throughout the 48 weeks of study (ITT observed analysis). Additionally, mean vRNA AAUCMB values were comparable between treatment groups at week 48: −2.41 log10 c/mL for the 908 BID group and −2.32 log10 c/mL for the NFV BID group. The mean treatment difference in AAUCMB was −0.082 (95% CI: −0.333 to 0.169) when adjusted for randomization strata. The upper boundary of the 95% CI was <0.5 log10 c/mL, indicating that 908 BID was noninferior to NFV BID with respect to this endpoint at week 48.

Back to Top | Article Outline

CD4 Cell Counts

In an ITT observed analysis, median CD4 cell counts showed similar increases for both groups over 48 weeks of treatment. Median changes from baseline at week 48 were 201 cells/mm3 for the 908 group and 216 cells/mm3 for the NFV group.

Back to Top | Article Outline

Safety Analyses

Adverse events were assessed in relation to the entire drug regimen, including the ABC and 3TC NRTI backbone. When considering drug-related adverse events of at least moderate severity (≥ grade 2), 30% (50/166) of patients in the 908 BID group and 34% (28/83) of patients in the NFV BID group reported at least one grade 2–4 drug-related adverse event. The most common drug-related grade 2–4 adverse events are presented in Table 3. The only grade 2–4 drug-related adverse event that was significantly different between the treatment groups was diarrhea (all grade 2 events), which was more common in the NFV BID group than the 908 BID group: 18% (15/83) of patients vs. 5% (8/166) of patients (P = 0.002). The median duration of this drug-related grade 2–4 diarrhea was also longer in the NFV BID group than the 908 BID group: 32 days vs. 10 days.



The incidence of drug-related grade 2–4 not otherwise specified rash was higher in the 908 BID group (7%, 12/166) than the NFV BID group (2%, 2/83), although this difference did not reach statistical significance. The incidence of grade 2–4 ABC hypersensitivity reaction (HSR) was similar between the groups (908 BID: 9%, 15/166; NFV BID: 5%, 4/83). Rash reported in conjunction with an abacavir HSR was entered into the database under abacavir HSR and was not entered or counted as the isolated event “rash.” A total of 14 subjects (6%) withdrew from the study due to an adverse event (908 BID: 5%, 9/166; NFV BID: 6%, 5/83 (Table 3). However, no single event leading to withdrawal from the study occurred in >2 patients in either treatment group.

The incidence of individual treatment-emergent grade 3/4 clinical chemistry abnormalities was generally low and comparable between treatment groups (Table 4). Elevations in liver enzymes were predominantly seen in patients entering the study with evidence of coinfection with hepatitis B or C. The incidence of grade 3/4 serum lipase elevations was higher in the 908 group than in the NFV group (8%, 13 patients, vs. 4%, 3 patients). None of these lipase elevations were clinically significant; all were asymptomatic, not linked to pancreatitis, occurred on <3 occasions per patient, and resolved without treatment or interruption of study drug.



No median fasting cholesterol values through 48 weeks met levels where clinical intervention would be recommended based on the NCEP guidelines. 9 For the 908 BID and NFV BID groups, respectively, median changes from baseline in total cholesterol showed increases of 39.8 mg/dL and 36.2 mg/dL, HDL cholesterol increased 10.1 mg/dL and 4.8 mg/dL, and LDL cholesterol increased 24.0 mg/dL and 25.1 mg/dL, respectively. A slightly greater increase in triglycerides was observed in the NFV BID group compared with the 908 BID group (Fig. 3). At 48 weeks, 26/144 patients (18%) in the 908 group and 12/68 patients (18%) in the NFV had LDL levels ≥160 mg/dL, at which point intervention is recommended per the NCEP guidelines for patients with few baseline risk factors. There was no appreciable change in the ratio of total cholesterol/HDL cholesterol during the course of the study in either treatment group (Fig. 4). These data are supported by a shift analysis of total cholesterol/HDL ratios, which revealed that for the 908 BID group, 86% (94/109) of the patients who provided fasting lipid data and who had ratio values below the NCEP cutoff (≤6.5) at baseline maintained their ratio below this level through the week 48 visit. For the NFV BID group, 78% (31/40) of the patients who provided fasting lipid data had ratio values below the NCEP cutoff (≤6.5) at baseline and maintained this ratio below threshold through week 48. Normalization of HDL cholesterol was achieved more frequently in the 908 BID group than the NFV BID group. In the 908 group, 6 subjects (4%) reported use of antihyperlipidemic medication during the trial, 5 of whom started the medication during the study. In the NFV BID group, all 3 subjects (4%) who took concomitant antihyperlipidemics started these medications during the study.





Back to Top | Article Outline


The patient population enrolled in this study was diverse in nature with respect to both gender and ethnicity and generally had more advanced HIV disease at entry than has been typically seen in studies of ART-naive populations. 10,11 This is reflected in the relatively high baseline viral loads, low CD4 cell counts, and the high proportion of patients entering this trial with history of a CDC class C event (Table 1). Notably, this population is more advanced than the study populations of recently published clinical trials in ART-naive HIV-infected patients. 12,13 Furthermore, the diversity of the population enrolled in this study is reflective of the current trends in recently reported HIV cases, with the incidence increasing in women and minorities, while remaining relatively stable in other demographic sectors. 8

APV30001 (NEAT) was the first of several phase 3 studies of 908, and the design considerations should not be overlooked. This study was conducted in ART-naive patients with the primary aim of generating a robust safety and tolerability database on 908. When the study was initiated and conducted, the comparator drug selected was NFV, this being the most widely used PI in this patient sector at that time as well as being included in the Department of Health and Human Services treatment guidelines as “strongly recommended” for first-line therapy use. 14

To gather adequate safety and tolerability information on 908, randomization was performed in a 2:1 ratio in favor of 908. One impact of this design is the limited size for the comparator group, such that results from this study should not be overinterpreted. However, to be able to make some comparisons between treatment groups, the study was powered on a secondary efficacy endpoint allowing for an analysis of the noninferiority of 908 BID to NFV BID.

When considering the proportion of subjects achieving undetectable levels of plasma HIV-1 RNA, the results of this study indicate that after 48 weeks of treatment, a greater proportion of patients in the 908 BID group than the NFV BID group achieved vRNA <400 c/mL (66 vs. 51%; ITT RD = F). While the study was not powered to assess a difference in proportions achieving undetectable levels of vRNA, it is noteworthy that the 95% CI was to the right of zero (2–28%), suggesting a true difference between the treatment groups in this study. In assessing this endpoint based on the ITT M = F population, the 95% CI (5–30%) also suggests a difference in favor of 908. Furthermore, these differences are maintained using the more stringent threshold of vRNA <50 c/mL for the ITT RD = F analysis (95% CI: 2–27%).

Perhaps the most notable observation in this study was the fact that 908 BID maintained antiviral activity across all screening vRNA stratification groups. Specifically at week 48, the proportion of 908 BID patients achieving vRNA <400 c/mL was similar for those subjects entering the study with screening vRNA either <100,000 or >100,000 c/mL (65 and 67%, ITT RD = F). In contrast, the response rates observed for NFV BID patients were not maintained and decreased in patients entering the study with high vRNA as noted by response rates of 63 and 35% for patients with screening vRNA <100,000 and >100,000 c/mL, respectively. Additionally, an analysis of patients who entered the study with vRNA >500,000 c/mL displayed continued maintenance of viral suppression in patients receiving 908 BID but showed a continuing trend in erosion of efficacy in those receiving NFV BID (65 and 31%, respectively).

While patients entering the study were not stratified based on CD4 cell counts, the fact that a considerable proportion of patients came into the study with low CD4 cells facilitated an analysis of response by this baseline immunologic parameter. While the 908 BID response was maintained in patients entering the study with CD4 cell counts < 200 cells/mm3 compared with those entering with CD4 cell counts >200 cells/mm3, there was a decrease in response rate for 908 BID within the lowest CD4 cell count (<50 cells/mm3) group; however, this decrease in response was less than the magnitude of the decrease seen with NFV BID. Finally, it was observed that the NFV BID group generally showed a somewhat lower antiviral response rate at 48 weeks in patients with advanced HIV disease (based on either high vRNA or low CD4+ cell counts at entry). In contrast, as discussed earlier, the vRNA response rate at 48 weeks was maintained in the 908 BID group regardless of the baseline vRNA strata.

Patients entering the study had low median baseline CD4 cell counts (214 and 212 cells/mm3 for the 908 BID and NFV BID groups, respectively). However, after 48 weeks on study, the median CD4 cell count was >400 cells/mm3 in both arms, indicating significant immune reconstitution over the course of the study. The increase from baseline in CD4 cell counts at week 48 in the 908 BID and NFV BID groups was similar, and these increases appeared not to have plateaued, suggestive of the possibility for continued immune system recovery beyond 48 weeks of treatment.

The virologic success rate (vRNA <400 c/mL) in the NFV BID group at week 48 (51 and 48%, ITT RD = F and ITT M = F, respectively) was somewhat lower than that reported in previous studies of NFV in treatment-naive subjects. 13,15–17 It is unclear why this somewhat lower response with NFV was observed in NEAT, although several factors may have contributed. First, 3 of the comparator studies were reported in clinical trials in which NFV was administered 3 times daily (AVANTI 3, BMS 007, and M98-863) rather than BID (BMS 008). Although therapeutic equivalence of BID and 3 times daily NFV dosing has been shown in other studies, 18–22 it is not known if this may have been a contributing factor in this study. Second, as stated previously, the patient population enrolled into this trial had more advanced HIV disease, a factor that may have contributed to the overall NFV response rate observed in this study. Finally, the NFV arm in the NEAT study was small (n = 83) relative to the aforementioned clinical trials, potentially representing a conservative estimate of response.

Both regimens were generally well tolerated. In the NFV BID group, gastrointestinal adverse events were more frequently reported than in the 908 BID group; drug-related grade 2–4 diarrhea was reported by 18% of subjects, which was consistent with the incidence reported in other studies evaluating NFV. 18 The difference between the 2 groups in the incidence of drug-related diarrhea (18 vs. 5%) was statistically significantly different (P = 0.002), and, furthermore, the diarrhea was of longer duration on NFV BID. Since this study was conducted, a new formulation of NFV has become available (625-mg tablets), potentially reducing the pill burden to 2 pills BID. However, some reports indicate that the incidence of diarrhea in patients receiving the newly Food and Drug Administration–approved 625-mg formulation may exceed that reported with the original NFV tablet presentation. 23

Drug hypersensitivity to ABC was the most common grade 2–4 drug-related adverse event in the 908 BID group (9% of patients), followed by rash (7%). These incidences noted here were not statistically significantly different from the incidences observed in patients receiving NFV. Few subjects in either group discontinued treatment due to adverse events, with no differences seen between treatment groups.

The overall incidence of subjects discontinuing ABC for possible HSR (all grades) was similar in both treatment groups (908 BID: 10%, 16/166; NFV BID: 7%, 6/83). Two subjects in the 908 BID group withdrew from the study due to possible abacavir HSR; the remaining subjects in both treatment groups exchanged ABC for another NRTI and continued in the study. The incidence of possible abacavir HSR reported in this study is higher than the approximate 5% previously reported in clinical trials. 24 This is consistent with a heightened awareness of abacavir HSR, and treating physicians are adopting an appropriately conservative approach to the diagnosis of abacavir HSR. 25,26

When considering laboratory changes, the median values of the following clinical laboratory parameters were within normal limits at week 48: triglycerides, total cholesterol, HDL cholesterol, LDL cholesterol, ALT, AST, and lipase. The 908 BID group had a smaller median increase from baseline in triglycerides and a larger, desirable median increase in HDL cholesterol than was observed in the NFV BID group. Additionally, the proportion of subjects receiving 908 BID who had undesirable HDL levels (<40 mg/dL) decreased over 48 weeks. The favorable lipid profile observed in this study was expected since previous in vivo work has shown that amprenavir has little or no effect on lipogenesis or lipid metabolism relative to other marketed PIs. 27,28

GW433908 offers several potential advantages over NFV with respect to tolerability, although with the introduction of the new 625-mg tablet presentation of NFV, the pill burden is reduced. However, the impact of this new formulation on clinical safety and efficacy is unknown, albeit reference to increased rates of diarrhea have been cited and it remains to be seen whether there is an overall favorable impact on the tolerability profile of NFV. In this study, gastrointestinal adverse events, common with PIs, were less common and less severe with 908 BID. When combined with a low pill burden, BID or QD prescribing flexibility, and the absence of food and fluid restrictions, 908 demonstrates several real-world benefits that should be considered when judging the effect of ART regimens on quality of life.

Resistance data from this study have previously been published in part 29 and will be the subject of a separate manuscript. To summarize, in those subjects who developed protease mutations, the mutations selected were consistent with the known amprenavir profile and included I54L/M, V32I, I47V, and M46I. Continuation of 908 BID therapy in the face of ongoing viral replication resulted in the evolution of the I54L genotype to the amprenavir-associated I50V substitution in virus from 1 subject. Little or no cross-resistance to other PIs was conferred by the protease mutations selected by 908. Furthermore, mutations observed with other PIs (D30N, I54V, V82A/T/S, L90M) were not observed following treatment with 908 BID. The spectrum of mutations selected by 908 is consistent with that observed with the active moiety, APV, 30 and is distinct from that observed with most other PIs. Subjects experiencing virologic failure on NFV developed protease mutations, selecting the common NFV resistance–associated mutations including D30N and L90M.

In conclusion, the patient population in this study was diverse, representative of patients with relatively advanced HIV disease and immune deficiency, and generally reflective of the changing demographics of the HIV epidemic. GW433908 demonstrated a potent and durable antiviral effect as shown by the proportion of patients achieving <400 and <50 copies/mL, by the substantial immunologic responses, and by the durability of response through 48 weeks of treatment. Furthermore, 908 was generally well tolerated in this treatment-naive patient population. Taken together, these efficacy, safety, tolerability, and resistance data demonstrate the utility of 908 as an early treatment option for persons living with HIV.

Back to Top | Article Outline


The following investigators participated in this study: J. Boghossian, MD, Newark, NJ; M. Borucki, MD, Galveston, TX; L. Brown, MD, Brooklyn, NY; A. Brutus, MD, Brooklyn, NY; D. Buffington, PharmD, Tampa, FL; J. DeSimone, MD, Philadelphia, PA; R. Fetchick, MD, San Antonio, TX; G. Frechette, MD, New York, NY; G. Gray, MD, Soweto, Republic of South Africa; H. Grossman, MD, New York, NY; W. Gaman, MD, Irving, TX; J. Giron, MD, Orlando, FL; W. Harley, MD, Charlotte, NC; W. Khayr, MD, N. Chicago, IL; I. Lowy, MD, New York, NY; G. Martinez, MD, Cotto Laurel, Puerto Rico; J. Nadler, MD, Tampa, FL; R. Pollard, MD, Galveston, TX; A. Quinones, MD, Fort Lauderdale, FL; R. Reichman, MD, Rochester, NY; A. Rodriguez-French, MD, Panama City, Panama; J. Santana, MD, Santurce, Puerto Rico; G. Sepulveda, MD, Ponce, Puerto Rico; S. Shah, MD, Houston, TX; D. Stein, MD, Bronx, NY; C. Steinhart, MD, Miami, FL; L. Tkatch, MD, Harrisburg, PA; C. Walworth, MD, Newport Beach, CA; N. Zide, MD, Hollywood, CA; C. Zurawski, MD, Atlanta, GA.

The authors thank Justin Cook for editing and writing assistance during preparation of the manuscript.

Back to Top | Article Outline


1. Falcoz C, Jenkins JM, Bye C, et al. Pharmacokinetics of GW433908, a prodrug of APV, in healthy male volunteers. J Clin Pharmacol. 2002; 42:887–898.
2. Wood R, Arasteh K, Stellbrink H-J, et al. A 6-week randomised controlled trial to compare the tolerability, pharmacokinetics and antiviral activity of GW433908 and amprenavir in HIV-1-infected patients. 2003;(In Press)
3. Wire MB, Preston SL, Ballow C, et al. An assessment of plasma amprenavir (APV) pharmacokinetics (PK) following administration of two GW433908 (908) and ritonavir (RTV) QD regimens in combination with efavirenz (EFV) in healthy adult subjects (APV10009) [1734]. Paper presented at: 41st Interscience Conference on Antimicrobial Agents and Chemotherapy; December 16–19, 2001; Chicago.
4. Wire MB, Ballow C, Preston S, et al. Pharmacokinetics and safety of GW433908 and ritonavir, with and without efavirenz, in healthy volunteers. 2003;(Submitted)
5. Maguire M, Shortino D, Klein A, et al. Emergence of resistance to protease inhibitor amprenavir in human immunodeficiency virus type 1-infected patients: selection of four alternative viral protease genotypes and influence of viral susceptibility to coadministered reverse transcriptase nucleoside inhibitors. Antimicrob Agent Chemother. 2002; 46:731–738.
6. Maguire M, Guinea R, Griffin P, et al. Human immunodeficiency virus type 1 p1/p6 protease cleavage site changes at Gag positions L449 or P453 are linked to I50V protease mutants in vivo and cause reduction of sensitivity to amprenavir and improved viral fitness. J Virol. 2002; 76:7398–7406.
7. Data on file. Research Triangle Park, NC: GlaxoSmithKline.
8. Centers for Disease Control and Prevention (CDC). HIV/AIDS Surveillance Report: US HIV and AIDS Cases Reported through December, 2000. 12:1–44. Available at:
9. Executive summary of the third report of the National Cholesterol Education Program (NCEP) expert panel on detection, evaluation, and treatment of high blood cholesterol in adults (Adult Treatment Panel III). JAMA. 2001; 285; 19:2486–2497.
10. Armstrong W, Calabrese L, Taege A. HIV update 2002: delaying treatment to curb rising resistance. Cleve Clin J Med. 2002; 69:995–999.
11. US Department of Health and Human Services. Guidelines for the use of antiretroviral agents in HIV-infected adults and adolescents. Available at: Accessed May 1, 2003.
12. Squires KE, Thiry A, Giordano M, et al. 2002. Atazanavir (ATV) qd and efavirenz (EFV) qd with fixed-dose ZDV+3TC: comparison of antiviral efficacy and safety through wk 48 (AI424-034) [H-1076]. Paper presented at: 42nd Interscience Conference on Antimicrobial Agents and Chemotherapy; September 27–30, 2002; San Diego.
13. Walmsley S, Bernstein B, King M, et al. Lopinavir-ritonavir versus nelfinavir for the initial treatment of HIV infection. N Engl J Med. 2002; 346:2039–2046.
14. US Department of Health and Human Services. Guidelines for the use of antiretroviral agents in HIV-infected adults and adolescents. Available at: Accessed January 28, 2000.
15. Johnson M, Beall G, Badley A, et al. ABT-378/ritonavir (ABT-378/R) versus NFV in antiretroviral naive subjects: week 48 comparison in a phase III blinded randomized clinical trial. AIDS. 2000; 14(suppl 4):S7.
16. Piliero PJ, Cahn P, Pantaleo G, et al. Atazanavir: a once-daily protease inhibitor with a superior lipid profile: results of clinical trials at week 48 [706-T]. Paper presented at: 9th Conference on Retroviruses and Opportunistic Infections; February 24–28, 2002; Seattle.
17. Gartland M on behalf of the AVANTI Study Group. AVANTI 3: a randomized, double-blind trial to compare the efficacy and safety of lamivudine plus zidovudine versus lamivudine plus zidovudine plus NFV in HIV-1-infected antiretroviral-naive patients. Antivir Ther. 2001; 6:127–134.
18. Clark R, Currier JS, Greiger P, et al. Efficacy and safety of BID and TID regimens of NFV (NFV) + stavudine (d4T) + lamivudine (3TC) in HIV-infected women: 48-week results of the Women First Study [1988]. Paper presented at: 39th Interscience Conference on Antimicrobial Agents and Chemotherapy; September 26–29, 1999; San Francisco.
19. Munsliff AV, Brutus A, Elizee M. Switching dosing of NFV from TID to BID maintains virological efficacy. Antivir Ther. 1998; 3(suppl 5):51. Abstract 51.
20. Pedersen A, Antunes F, Arasteh KN, et al. A comparison of the long-term antiviral efficacy of bid and tid dosing of NFV in combination with stavudine (d4T) and lamivudine (3TC) beyond 48 weeks. Paper presented at: 7th European Conference on Clinical Aspects and Treatment of HIV Infection; October 23–27, 1999; Lisbon.
21. Johnson M, Nelson M, Peters B, et al. A comparison of BID and TID dosing of NFV mesylate when given in combination with stavudine (d4T) and lamivudine (3TC) for up to 48 weeks [I216]. Paper presented at: 38th Interscience Conference on Antimicrobial Agents and Chemotherapy; September 24–27, 1998; San Diego.
22. Bardsley-Elliot A, Plosker GL. NFV: an update on its use in HIV infection. Drugs. 2000; 59:581–620.
23. Viracept (nelfinavir mesylate) [Product information]. La Jolla, CA: Agouron Pharmaceuticals: April 2003.
24. Hetherington S, McGuirk S, Powell G, et al. Hypersensitivity reactions during therapy with the nucleoside reverse transcriptase inhibitor abacavir. Clin Ther. 2001; 23:1603–1614.
25. Hernandez J, Cutrell A, Bonny T, et al. Diagnosis of abacavir hypersensitivity reactions among patients not receiving abacavir in two blinded studies. Paper presented at the 5th International Workshop on Adverse Drug Reactions and Lipodystrophy in HIV; July 8–11, 2003; Paris, France. Abstract 134.
26. Cutrell A, Hernandez J, Edwards M, et al. Clinical risk factor analysis of hypersensitivity reactions to abacavir: retrospective analysis of over 8,000 subjects receiving abacavir in 34 clinical trials. Paper presented at: 43rd Annual ICAAC; September 14–17, 2003; Chicago, IL. Abstract H-2013)
27. Fung HB, Kirschenbaum HL, Hameed R. Amprenavir: a new human immunodeficiency virus type 1 protease inhibitor. Clin Ther. 2000; 22:549–572.
28. Noble S, Goa KL. Amprenavir: a review of its clinical potential in patients with HIV infection. Drugs. 2000; 60:1383–1410.
29. MacManus S, Yates PJ, White S, et al. GW433908 in ART-naive subjects: absence of resistance at 48 weeks with boosted regimen and APV-like resistance profile with unboosted regimen [598]. Paper presented at: 10th Conference on Retroviruses and Opportunistic Infections; February 10–14, 2003; Boston.
30. International AIDS Society-USA. Drug resistance in HIV-1 update. Top HIV Med. 2002; 10:11–15.

HIV; GW433908; nelfinavir; NEAT

© 2004 Lippincott Williams & Wilkins, Inc.