|
|
|
|
|
Session 101
Poster Abstracts Antiretroviral Therapy: Regimens, Predictors of Response, and Clinical Outcomes Thursday, 1:30 - 3:30 pm Hall A |
Background: In the HAART
era, few data compare rates of death according to initial drug regimen. We
examined whether rates of death differed according to initial treatment regimen
among HIV+ persons starting antiretroviral therapy with a non-nucleoside
reverse transcriptase inhibitor (NNRTI), protease inhibitor (PI), or triple
nucleoside-based regimen.
Methods: The ART Cohort
Collaboration is a multinational cohort study of antiretroviral naïve patients
initiating HAART. Cox models were used to estimate hazard ratios comparing 8
third drugs: efavirenz
(EFV) (comparator), nevirapine (NVP), nelfinavir (NFV), indinavir (IDV),
ritonavir (RTV), RTV-boosted PI (amprenavir
[APV], lopinavir [LPV], saquinavir
[SQV], and IDV), SQV (soft and hard), and abacavir [ABC].
We also compared NRTI pairs: zidovudine (AZT)/lamivudine (3TC)
(comparator), stavudine (d4T)/3TC, d4T/didanosine (ddI), and others. Multivariable models were adjusted for
age, gender, transmission risk group, previous AIDS and pre-HAART CD4 count,
and viral load. In addition, the effect of third drug was adjusted for NRTI
pair and vice-versa.
Results:
Analyses
were based on 17,666 patients with 55,622 person years at risk and 895 deaths.
Median follow-up was 3.2 years, maximum 7.5 years. The most common third drugs
were IDV (26%), NFV (22%), EFV (13%), and NVP (13%). The most common nucleoside
pairs were AZT/3TC (62%), d4T/3TC (20%), and d4T/ddI (10%). There was little
evidence of between cohort heterogeneity, but some
evidence that death rates varied according to the third drug (p = 0.08) and nucleoside pair (p = 0.006). As shown below the adjusted hazard
ratios were 1.65 (1.16 to 2.36) for NVP vs EFV and
1.35 (1.14 to 1.59) for d4T/3TC vs AZT/3TC.
|
|
Patients |
Deaths |
Crude HR |
Adjusted
HR |
|
EFV |
2,254 |
58 |
1 |
1 |
|
NVP |
2,378 |
92 |
1.41 (1.00, 2.00) |
1.65 (1.16, 2.36) |
|
NFV |
3,802 |
191 |
1.38 (1.00, 1.89) |
1.07 (0.77, 1.47) |
|
RTV |
1,148 |
91 |
1.58 (1.06, 2.35) |
1.36 (0.91, 2.02) |
|
SQV |
1,145 |
67 |
1.15 (0.75, 1.75) |
1.15 (0.75, 1.75) |
|
RTV-boosted PI |
1,285 |
52 |
1.57 (1.07, 2.32) |
1.14 (0.77, 1.69) |
|
IDV |
4,555 |
318 |
1.47 (1.04, 2.08) |
1.20 (0.85, 1.71) |
|
ABC |
1,099 |
26 |
1.07 (0.66, 1.74) |
1.27 (0.77, 2.09) |
|
|
|
|
|
|
|
AZT/3TC |
10,964 |
462 |
1 |
1 |
|
d4T/3TC |
3,468 |
255 |
1.60 (1.36, 1.88) |
1.35 (1.14, 1.59) |
|
d4T/ddI |
1,716 |
96 |
1.25 (0.99, 1.57) |
1.15 (0.91, 1.45) |
|
Other |
1,518 |
82 |
1.10 (0.86, 1.41) |
1.13 (0.88, 1.46) |
Conclusions: We
generally did not find large differences in mortality rates according to
initial HAART regimen. The observed between-regimen variability in prognosis
could be due to unmeasured confounding by indication rather than to differences
in efficacy or adverse event profile between regimens.
Keywords: Antiretroviral naive; HAART; Survival
 |