Transcript Long-Term Comparison of Nevirapine Versus Efavirenz

Long-Term Comparison of Nevirapine Versus Efavirenz When Combined with Other Antiretroviral Drugs in
HIV-1 Positive Antiretroviral-Naïve Persons- The NNRTI Substudy of the CPCRA 058 FIRST Study
van den Berg-Wolf M1, Peng G2, Xiang Y2, Huppler Hullsiek K2, Chen L2, MacArthur RD3, Novak RM4, Kozal M5, Schmetter B6, Henely C7, Dehlinger M8
1. Temple University, Philadelphia, Pa, USA 2. University of Minnesota, Minneapolis, Mn, USA, 3. Wayne State University, Detroit, Mi, USA 4. University of Illinois, Chicago, Il, USA
5. Yale University School of Medicine, New Haven, Ct, USA 6. Social and Scientific Systems, Silver Springs, Md, USA 7. Meridian Health System, Neptune, N J, USA 8. Division of AIDS, NIAID, Bethesda, Md, USA
BACKGROUND
TABLE 3: Antiretroviral Drug Changes* During Follow-up
CPCRA 058 (FIRST) was a randomized open-label trial comparing three initial antiretroviral treatment strategies in
1,397 HIV-infected antiretroviral-naïve persons. Treatment strategies are defined by classes of drugs, not specific
antiretrovirals (ARVs). Enrollment was from March 1999 to January 2002 at 17 clinical units in the United States.
Before randomization in the main study, participants were given the option of additional substudy randomizations or
pre-selecting the drugs within each strategy arm. Participants consenting to the NNRTI substudy were further
randomized to efavirenz or nevirapine (1:1 allocation), taken together with other classes of drugs within the strategy
(the randomized cohort; Figure 1). In addition, we report on participants who were randomized to the NNRTI containing
arms of FIRST, but chose to select the NNRTI used in combination with other ARVs (the non-randomized cohort; Figure
1).
METHODS
Efavirenz (N=111)
Nevirapine (N=117)
Switched to alternate NNRTI
3.6
2.6
Discontinued assigned NNRTI
69.4
72.6
Switched to alternate NNRTI
1.9
3.4
Discontinued assigned NNRTI
69.2
70.7
Initiated PI
28.8
44.8
Switched NRTI
76.9
74.1
Switched to alternate NNRTI
5.1
1.7
Discontinued assigned NNRTI
69.5
74.6
Discontinued PI
62.7
61.0
Switched PI
11.9
25.4
Switched NRTI
74.6
84.7
Overall(%)
2-class strategy (%)
3-class strategy (%)
• Population: participants naïve to active combination antiretroviral treatment.
• Primary Endpoint: time to HIV-RNA > 50 copies/mL (at or after month 8 visit) or death.
• Secondary Endpoints: progression to AIDS or death, changes in CD4+ cell count, percent of participants with
HIV-RNA < 50 copies/mL, HIV drug resistance* seen at the first time HIV-RNA > 1000 copies/mL (at or after month 4
visit), safety and tolerability of the drugs, and grade 4 adverse events.
• ARV therapy management: participants enrolled into the FIRST study could change ARVs for any reason;
*Categories are not mutually exclusive. There were no significant differences between the groups.
however, they were encouraged to stay within the assigned strategy if possible.
TABLE 4: Event Rates
• Sample size: 200 primary events were required for 80% power to detect a hazard ratio of 0.67 at the 0.05 level of
Efavirenz (N=111)
significance (2-sided). When the FIRST study closed, 187 primary events had occurred within the NNRTI substudy.
• Statistical Methods (intent to treat):
• Time to event analyses: Kaplan-Meier life table summaries and Cox’s proportional hazards models. The models
were stratified by randomized strategy in the main FIRST study (2-class strategy with no PI or 3-class strategy).
Hazard ratios compare efavirenz to nevirapine. Rates are per 100 person years of follow-up.
• Changes in CD4+ cell count: longitudinal regression models adjusted for baseline values.
• Proportion with HIV-RNA < 50 copies/mL: models for longitudinal binary data.
• Sensitivity Analyses:
• For both the randomized cohort and the non-randomized cohort, analyses were repeated with models adjusting
for baseline factors considered to be potential confounding variables (age, race, gender, prior AIDS, history of
injection drug use, baseline CD4 cell count and HIV RNA level).
• Those models were repeated for the combined data set, including an indicator for randomization status and the
interaction between randomization status and NNRTI drug.
*Resistance was defined as definite drug resistance by genotype (TRUGENE HIV-1 Genotyping Kit and CPCRA
interpretive algorithm v4.0).
Figure 1
Randomized to FIRST Study
PI Strategy
(PI + NRTIs)
N=470
NNRTI Strategy
(NNRTI + NRTIs)
N=463
NNRTI
Prescribed
substudy
Nevirapine
Randomization (N=111)
Nevirapine
(N=58)
Nevirapine
(N=59)
Efavirenz
(N=52)
Randomized to
NNRTI substudy
Nevirapine
58 + 59 = 117
Rate1 Hazard Ratio2 95% CI) P-value
No. Event
Rate1
No. Event
Primary endpoint3
89
41.2
98
42.8
0.92 (0.69-1.23)
0.59
Death
20
3.8
18
3.2
1.18 (0.62-2.23)
0.61
HIV RNA > 50
copies/mL
82
42.0
94
45.9
0.89 (0.66-1.19)
0.42
AIDS or death
34
7.4
23
4.4
1.67 (0.98-2.83)
0.06
HIV RNA > 1000
copies/mL
71
28.7
86
36.4
0.80 (0.58-1.10)
0.17
With resistance to
NNRTI drugs
32
12.9
49
20.6
0.65 (0.41-1.01)
0.05
NRTI drugs
5
2.0
25
10.5
0.20 (0.08-0.52)
<0.01
Any ARV drugs
33
13.3
54
22.7
0.60 (0.39-0.93)
0.02
Grade 4 event
24
5.4
43
10.2
0.55 (0.33-0.9.)
0.02
Discontinue
randomized drug
due to toxicity
20
4.5
16
3.2
1.34 (0.70-2.59)
0.38
1. Rates are per 100 person-years
2. Hazard ratios are for the comparison of efavirenz versus nevirapine
3. Time to death or to HIV RNA level > 50 copies/mL (at or after month 8 visit)
3-Class Strategy
(PI + NNRTI + NRTIs)
N=464
Prescribed
NNRTI
Prescribed
Efavirenz
substudy
Nevirapine
(N=237) Randomization (N=150)
Nevirapine (N=117)
Prescribed
Efavirenz
(N=179)
Figure 2
Time to Death or First HIV RNA > 50 Copies/mL
at or After Month 8 Visit
Efavirenz
(N=59)
Figure 3
Percent with HIV RNA < 50 Copies/mL
Not Randomized to
NNRTI substudy
Efavirenz
52 + 59 = 111
Nevirapine
111 + 150 = 261
Efavirenz
237 + 179 = 416
RESULTS for Randomized Cohort
• The NNRTI substudy enrolled 228 participants: 111 to the efavirenz group and 117 to the nevirapine group (Figure
1). An additional 416 participants were prescribed efavirenz, and 261 were prescribed nevirapine.
• Median follow-up time was 65 months (inter-quartile range 57–71 months).
• Among participants randomized to the 3-class strategy, nelfinavir was the most frequently prescribed PI. One person
in the nevirapine group was prescribed efavirenz at study entry (Table 2).
• About 75% discontinued the assigned NNRTI drug during follow-up (p=0.68 for difference between the two groups;
Table 3).
• More participants in the nevirapine group either initiated a PI (among those in the 2-class strategy, p=0.08) or
switched PIs (among those in the 3-class strategy, p=0.06) than the efavirenz group (Table 3).
Figure 4
Mean CD4+ Cell Count Change from
Baseline
TABLE 5: Hepatic Toxicity
Efavirenz
(N=111)
No.
Events
%
No.
Events
%
Pvalue
>10 ULN* at least
once during follow-up
or grade 4 AE
7
6.3
10
8.5
0.52
> 5 ULN* at least
once during follow-up
16
14.4
18
15.4
0.84
>2.5 ULN* at least
once during follow-up
30
27.0
38
32.5
0.37
• There was no significant difference between the groups for the primary endpoint (time to death or HIV RNA level >
50 copies/mL): hazard ratio (HR) = 0.92, p=0.59; Figure 2 and Table 4).
• The rate of AIDS or death was higher in the efavirenz group (HR = 1.67, p=0.06; Table 4); The rate of death only did
not differ between the groups.
• The rate of grade 4 events was lower for the efavirenz group than for the nevirapine group (HR=0.55, p=0.02; Table
4). However, the percents with grade 4 elevated AST or ALT levels were similar (Table 5).
• The rate of discontinuation of randomized drugs due to toxicity was similar (HR=1.34, p=0.38, Table 4). The majority
(86%) of the toxicities were grade 1 to 3.
Nevirapine
(N=117)
* Upper limit of normal for AST or ALT, required at each follow-up visit
• The rates of virologic failure (HIV RNA > 1000 copies/mL) were similar (HR = 0.80, p = 0.17; Table 4). However, the
rate of virologic failure associated with any drug resistance was lower in the efavirenz group (HR = 0.60, p=0.02).
• The percent with HIV RNA < 50 copies/mL was not significantly different between the 2 groups (odds ratio = 1.26,
p=0.24; Figure 3).
• CD4+ cell increases from baseline did not differ significantly over follow-up (p=0.25; Figure 4). The average CD4+
cell count increases over the entire follow-up period were 172 and 153 cells/mm³ for the efavirenz and nevirapine
groups, respectively.
TABLE 1: Baseline Characteristics*
Efavirenz
(N=111)
Nevirapine
N=117)
Overall
(N=228)
Age (mean years)
38
36
37
Female (%)
23
22
23
Latino/Latina
16
18
17
African American
60
60
60
White/other
24
22
23
Prior AIDS event (%)
38
37
37
Hepatitis B (%)
5
5
5
Hepatitis C (%)
21
16
18
History of injection drug use (%)
15
18
17
2-class NNRTI Strategy (%)
47
50
48
CD4+ cells (median cells/mm3)
181
196
186
HIV RNA (median log
5.0
5.1
5.0
Race/Ethnicity (%)
10
copies/mL)
*There were no significant differences between the two groups.
Nevirapine
(N=117)
Randomized to Substudy
Unadjusted
Adjusted3
Not Randomized to
Substudy3
Primary2
0.92
(0.69-1.23)
0.95
(0.71-1.28)
0.79
(0.67-0.95)
0.79
(0.66-0.94)
0.30
AIDS or death
1.67
(0.98-2.83)
2.21
(1.27-3.84)
0.90
(0.64-1.26)
0.88
(0.63-1.23)
0.02
HIV RNA > 1000
copies/mL with
resistance
0.60
(0.39-0.93)
0.61
(0.39-0.95)
0.75
(0.55-1.02)
0.74
(0.54-1.01)
0.53
Grade 4 Event
0.55
(0.33-0.90)
0.51
(0.30-0.85)
0.89
(0.67-1.17)
0.88
(0.67-1.16)
0.08
Event
Combined
Data3
Interaction
P-value4
1 Hazard ratios (95% CI) are for the comparison of efavirenz versus nevirapine
2 Time to death or to HIV RNA level > 50 copies/mL (at or after month 8 visit)
3 Adjusted for age, gender, race, prior AIDS, history of injection drug use, baseline CD4 cell count and HIV RNA level
4 From an interaction test between study drug (efavirenz or nevirapine) and substudy randomization status
RESULTS of Comparison of Randomized and
Non-randomized Cohorts in FIRST
• The hazard ratios for the primary endpoint (HIV RNA > 50 copies/mL or death) favored the efavirenz group over the
nevirapine group for both the randomized and non-randomized cohorts (pooled HR = 0.79, 95% CI: 0.66-0.94; Table 6).
• The hazard ratios for virologic failure associated with any drug resistance also favored the efavirenz group. (pooled HR
= 0.74, 95% CI: 0.54-1.01; Table 6).
TABLE 2: Antiretroviral Drugs Prescribed
at Study Entry*
Efavirenz
(N=111)
TABLE 6: Comparison of Hazard Ratios1 of
Randomized and Non-randomized Cohorts in FIRST
Overall
(N=228)
PI(%)
• For AIDS or death events, there was a significant interaction between NNRTI drug and substudy randomization status
(Table 6). In the randomized substudy there was a significantly increased risk of AIDS or death in the efavirenz group
that was not present in the non-randomized cohort.
• In the combined cohort, there was no difference between the groups for the rate of grade 4 events (Table 6)
Nelfinavir
27
26
26
Indinavir
14
9
11
Ritonavir-boosted PI
14
15
14
No PI
46
50
48
Abacavir + Lamivudine
27
30
29
Didanosine + Stavudine
25
29
27
• In the NNRTI substudy, the efavirenz and nevirapine groups did not differ significantly for rate of HIV RNA > 50
copies/mL or death; however pooled results indicate that the nevirapine group has increased risk of the event.
Zidovudine + Lamivudine
22
16
19
• There was no differences found for CD4+ cell increases over time in the two groups.
Stavudine + Lamivudine
11
12
11
• Nevirapine was associated with more drug resistance at the time of virologic failure.
Single NRTI
12
9
11
Other
3
4
3
• The different results for AIDS or death events in the randomized and non-randomized cohorts suggest residual
confounding in the non-randomized cohort.
NRTI (%)
*One person randomized to the nevirapine group was prescribed efavirenz at study entry.
We would like to thank all the participants who volunteered to be part of this study.
Supported by grants (5U01AI042170-10 and 5U01AI046362-03) from the National Institute of Allergy
and Infectious Diseases, National Institutes of Health, USA. Reprints: [email protected]
• All results were consistent with on-treatment analyses.
CONCLUSIONS
• Larger, longer-term randomized trials are needed to assess the clinical outcome of AIDS or death with adequate
power comparing efavirenz with nevirapine as initial antiretroviral treatment options.
• Joint analyses of randomized and non-randomized cohorts can potentially improve power for treatment
comparisons and permit assessment of residual confounding for non-randomized analyses.